1<!DOCTYPE html><html lang="en"><head><meta charSet="utf-8" data-next-head=""/><title data-next-head="">Blog | Agentman</title><meta name="title" content="Blog | Agentman" data-next-head=""/><meta name="description" content="Explore the latest trends, insights, and best practices in AI agents, no-code development, and the future of business automation." data-next-head=""/><meta name="viewport" content="width=device-width, initial-scale=1" data-next-head=""/><meta name="keywords" content="agentic healthcare, AI agents, healthcare RCM, revenue cycle management, eligibility verification, prior authorization, inbox triage, AI voice agent, HIPAA compliant AI" data-next-head=""/><meta property="og:type" content="website" data-next-head=""/><meta property="og:title" content="Blog | Agentman" data-next-head=""/><meta property="og:description" content="Explore the latest trends, insights, and best practices in AI agents, no-code development, and the future of business automation." data-next-head=""/><meta property="og:url" content="https://agentman.ai/blog" data-next-head=""/><meta property="og:image" content="https://agentman.ai/og-image.jpg" data-next-head=""/><meta property="og:image:width" content="1200" data-next-head=""/><meta property="og:image:height" content="630" data-next-head=""/><meta property="og:site_name" content="Agentman" data-next-head=""/><meta name="twitter:card" content="summary_large_image" data-next-head=""/><meta name="twitter:title" content="Blog | Agentman" data-next-head=""/><meta name="twitter:description" content="Explore the latest trends, insights, and best practices in AI agents, no-code development, and the future of business automation." data-next-head=""/><meta name="twitter:image" content="https://agentman.ai/twitter-card-image.jpg" data-next-head=""/><link rel="canonical" href="https://agentman.ai/blog" data-next-head=""/><meta name="robots" content="index, follow" data-next-head=""/><meta name="author" content="Agentman" data-next-head=""/><link rel="preload" as="image" imageSrcSet="/_next/image?url=%2Fagentman-mark.svg&w=48&q=75 1x, /_next/image?url=%2Fagentman-mark.svg&w=96&q=75 2x" data-next-head=""/><link rel="preload" as="image" imageSrcSet="/_next/image?url=%2Fimages%2Fblog%2FNews7_image1.png&w=640&q=75 640w, /_next/image?url=%2Fimages%2Fblog%2FNews7_image1.png&w=750&q=75 750w, /_next/image?url=%2Fimages%2Fblog%2FNews7_image1.png&w=828&q=75 828w, /_next/image?url=%2Fimages%2Fblog%2FNews7_image1.png&w=1080&q=75 1080w, /_next/image?url=%2Fimages%2Fblog%2FNews7_image1.png&w=1200&q=75 1200w, /_next/image?url=%2Fimages%2Fblog%2FNews7_image1.png&w=1920&q=75 1920w, /_next/image?url=%2Fimages%2Fblog%2FNews7_image1.png&w=2048&q=75 2048w, /_next/image?url=%2Fimages%2Fblog%2FNews7_image1.png&w=3840&q=75 3840w" imageSizes="(max-width: 768px) 100vw, 580px" data-next-head=""/><link rel="icon" type="image/svg+xml" href="/favicon.svg"/><link rel="icon" href="/favicon.ico"/><link rel="icon" type="image/png" sizes="16x16" href="/favicon-16x16.png"/><link rel="icon" type="image/png" sizes="32x32" href="/favicon-32x32.png"/><link rel="icon" type="image/png" sizes="96x96" href="/favicon-96x96.png"/><link rel="apple-touch-icon" sizes="180x180" href="/apple-touch-icon.png"/><link rel="manifest" href="/site.webmanifest"/><link rel="preload" href="/fonts/geist-latin.woff2" as="font" type="font/woff2" crossorigin=""/><meta name="google-site-verification" content="q9vXaDu1EqFn-JzKOcVsuT6x3pYRhYobr-s6luoBizs"/><meta name="theme-color" content="#ffffff"/>
1<script type="application/ld+json">{"@context":"https://schema.org","@type":"Organization","name":"Agentman","url":"https://agentman.ai","logo":"https://agentman.ai/agentman-logo.png","sameAs":["https://www.linkedin.com/company/agentman"]}</script>
1<script>
vendor: 398 bytes, lines 1-9
1!function(f,b,e,v,n,t,s) 2{if(f.fbq)return;n=f.fbq=function(){n.callMethod? 3n.callMethod.apply(n,arguments):n.queue.push(arguments)}; 4if(!f._fbq)f._fbq=n;n.push=n;n.loaded=!0;n.version='2.0'; 5n.queue=[];t=b.createElement(e);t.async=!0; 6t.src=v;s=b.getElementsByTagName(e)[0]; 7s.parentNode.insertBefore(t,s)}(window, document,'script', 8'https://connect.facebook.net/en_US/fbevents.js'); 9fbq('init', '
91329073872051626
vendor: 29 bytes, lines 9-10
9'); 10fbq('track', 'PageView');
10</script>
10<noscript><img height="1" width="1" style="display:none" src="https://www.facebook.com/tr?id=1329073872051626&ev=PageView&noscript=1" alt=""/></noscript><link rel="preload" href="/_next/static/media/e4af272ccee01ff0-s.p.woff2" as="font" type="font/woff2" crossorigin="anonymous" data-next-font="size-adjust"/><link rel="preload" href="/_next/static/media/49eec060ce8bd0da-s.p.woff2" as="font" type="font/woff2" crossorigin="anonymous" data-next-font="size-adjust"/><link rel="preload" href="/_next/static/media/fa3e259cafa8f47e-s.p.woff2" as="font" type="font/woff2" crossorigin="anonymous" data-next-font="size-adjust"/><link rel="preload" href="/_next/static/media/eaead17c7dbfcd5d-s.p.woff2" as="font" type="font/woff2" crossorigin="anonymous" data-next-font="size-adjust"/><link rel="preload" href="/_next/static/css/600c51bac3fb070a.css" as="style"/><link rel="stylesheet" href="/_next/static/css/600c51bac3fb070a.css" data-n-g=""/><noscript data-n-css=""></noscript>
10<script defer="" nomodule="" src="/_next/static/chunks/polyfills-42372ed130431b0a.js"></script>
10<script src="/_next/static/chunks/webpack-3bf654b8151a69ab.js" defer=""></script>
10<script src="/_next/static/chunks/framework-b9fd9bcc3ecde907.js" defer=""></script>
10<script src="/_next/static/chunks/main-e77fbedafec9f4f2.js" defer=""></script>
10<script src="/_next/static/chunks/pages/_app-8d9a986492d03a57.js" defer=""></script>
10<script src="/_next/static/chunks/5171-a82b85c41efa5c38.js" defer=""></script>
10<script src="/_next/static/chunks/2749-dd6f885041d112c1.js" defer=""></script>
10<script src="/_next/static/chunks/8230-f7b73da2105800a2.js" defer=""></script>
10<script src="/_next/static/chunks/2517-347c27dff83c8124.js" defer=""></script>
10<script src="/_next/static/chunks/4372-591daab810a3553c.js" defer=""></script>
10<script src="/_next/static/chunks/2678-1b930249fcea6ec9.js" defer=""></script>
10<script src="/_next/static/chunks/7361-7075b16f4d8400d9.js" defer=""></script>
10<script src="/_next/static/chunks/7007-df33597f4553fa8a.js" defer=""></script>
10<script src="/_next/static/chunks/1448-8c32544325844c4a.js" defer=""></script>
10<script src="/_next/static/chunks/375-4c1adfca821bf5ce.js" defer=""></script>
10<script src="/_next/static/chunks/pages/blog-dce3e675275effec.js" defer=""></script>
10<script src="/_next/static/W-uwlKWK8YmOc7QuKgaOl/_buildManifest.js" defer=""></script>
10<script src="/_next/static/W-uwlKWK8YmOc7QuKgaOl/_ssgManifest.js" defer=""></script>
10</head><body><div id="__next"><div class="__variable_8b3a0b __variable_d501d1 __variable_0a80b4"><div class="min-h-screen"><main><nav class="sticky top-0 inset-x-0 z-50 bg-white border-b border-gray-200"><div class="relative max-w-7xl mx-auto transition-all duration-300 py-3"><div class="flex items-center justify-between px-4 sm:px-6"><div class="flex-shrink-0"><a class="group relative block" aria-label="Agentman home" href="/"><div class="relative flex items-center gap-2 transition-all duration-500 ease-out" tabindex="0"><img alt="Agentman" width="41" height="41" decoding="async" data-nimg="1" class="w-auto transition-all duration-300 h-8" style="color:transparent;object-fit:contain" srcSet="/_next/image?url=%2Fagentman-mark.svg&w=48&q=75 1x, /_next/image?url=%2Fagentman-mark.svg&w=96&q=75 2x" src="/_next/image?url=%2Fagentman-mark.svg&w=96&q=75"/><span class="font-bold leading-none tracking-tight text-charcoal-900 transition-all duration-300 text-2xl" style="font-family:Inter, system-ui, -apple-system, sans-serif">Agentman</span></div></a></div><div class="hidden lg:flex lg:items-center lg:space-x-2"><div class="relative"><button class="flex items-center gap-1 px-4 py-2 rounded-lg text-slate-600 font-medium transition-all duration-200 hover:text-charcoal-950 hover:bg-agentman-50">Products<svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-chevron-down h-4 w-4 transition-transform duration-200"><path d="m6 9 6 6 6-6"></path></svg></button></div><div class="relative"><button class="flex items-center gap-1 px-4 py-2 rounded-lg text-slate-600 font-medium transition-all duration-200 hover:text-charcoal-950 hover:bg-agentman-50">Industries<svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-chevron-down h-4 w-4 transition-transform duration-200"><path d="m6 9 6 6 6-6"></path></svg></button></div><div class="relative"><button class="flex items-center gap-1 px-4 py-2 rounded-lg text-slate-600 font-medium transition-all duration-200 hover:text-charcoal-950 hover:bg-agentman-50">Pricing<svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-chevron-down h-4 w-4 transition-transform duration-200"><path d="m6 9 6 6 6-6"></path></svg></button></div><a class="px-4 py-2 rounded-lg text-slate-600 font-medium transition-all duration-200 hover:text-charcoal-950 hover:bg-agentman-50" href="/blog">Blog</a><a class="px-4 py-2 rounded-lg text-slate-600 font-medium transition-all duration-200 hover:text-charcoal-950 hover:bg-agentman-50" href="/about">About</a><a class="px-4 py-2 rounded-lg text-slate-600 font-medium transition-all duration-200 hover:text-charcoal-950 hover:bg-agentman-50" href="/login">Login</a></div><div class="lg:hidden"><button class="inline-flex items-center justify-center p-2 rounded-lg text-charcoal-950 hover:text-agentman-600 hover:bg-agentman-50 focus:outline-none focus:ring-2 focus:ring-inset focus:ring-agentman-500 transition-all duration-200"><span class="sr-only">Open menu</span><div style="opacity:0;transform:rotate(90deg)"><svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-menu h-6 w-6"><line x1="4" x2="20" y1="12" y2="12"></line><line x1="4" x2="20" y1="6" y2="6"></line><line x1="4" x2="20" y1="18" y2="18"></line></svg></div></button></div></div><div class="pointer-events-none absolute inset-x-0 bottom-0 overflow-hidden"><div class="absolute bottom-0 left-0 h-[3px] bg-agentman-500" style="width:0%"></div></div></div></nav><main class="bg-white pt-10"><section class="max-w-[1200px] mx-auto px-6 pb-6"><div class="flex items-end justify-between gap-4"><div><h1 class="text-4xl sm:text-5xl font-bold text-charcoal-950 tracking-tight">Blog</h1><p class="mt-2 text-slate-600 text-lg flex items-center gap-3 flex-wrap">Please subscribe to Agentman's newsletters<a href="#newsletter" aria-label="Subscribe to Agentman's newsletters" class="inline-flex items-center gap-2 rounded-lg border border-agentman-200 bg-white px-4 py-2 text-sm font-semibold text-charcoal-950 no-underline shadow-sm hover:border-agentman-500 hover:bg-agentman-50 transition-colors"><img alt="" loading="lazy" width="22" height="24" decoding="async" data-nimg="1" class="h-5 w-auto shrink-0" style="color:transparent" srcSet="/_next/image?url=%2Fagentman-mark.svg&w=32&q=75 1x, /_next/image?url=%2Fagentman-mark.svg&w=48&q=75 2x" src="/_next/image?url=%2Fagentman-mark.svg&w=48&q=75"/>Subscribe</a></p></div><div class="flex items-center gap-3 flex-shrink-0"><div class="relative"><svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-search absolute left-3 top-1/2 -translate-y-1/2 w-4 h-4 text-slate-400 pointer-events-none"><circle cx="11" cy="11" r="8"></circle><path d="m21 21-4.3-4.3"></path></svg><input type="text" placeholder="Search posts..." aria-label="Search blog posts" class="w-48 sm:w-56 pl-9 pr-8 py-2 text-sm rounded-lg border border-slate-200 bg-white text-charcoal-950 placeholder:text-slate-400 focus:outline-none focus:ring-2 focus:ring-agentman-500 focus:border-transparent transition-colors" value=""/></div></div></div></section><section class="border-b border-slate-200"><div class="max-w-[1200px] mx-auto px-6"><div class="flex gap-6 overflow-x-auto scrollbar-hide -mb-px"><button class="flex items-center gap-1.5 pb-3 text-sm font-medium whitespace-nowrap border-b-2 transition-colors border-charcoal-950 text-charcoal-950">All<span class="text-xs px-1.5 py-0.5 rounded-full bg-charcoal-950 text-white">90</span></button><button class="flex items-center gap-1.5 pb-3 text-sm font-medium whitespace-nowrap border-b-2 transition-colors border-transparent text-slate-500 hover:text-slate-700">Newsletter<span class="text-xs px-1.5 py-0.5 rounded-full bg-slate-100 text-slate-500">7</span></button><button class="flex items-center gap-1.5 pb-3 text-sm font-medium whitespace-nowrap border-b-2 transition-colors border-transparent text-slate-500 hover:text-slate-700">Healthcare<span class="text-xs px-1.5 py-0.5 rounded-full bg-slate-100 text-slate-500">47</span></button><button class="flex items-center gap-1.5 pb-3 text-sm font-medium whitespace-nowrap border-b-2 transition-colors border-transparent text-slate-500 hover:text-slate-700">
10Platform<span class="text-xs px-1.5 py-0.5 rounded-full bg-slate-100 text-slate-500">24</span></button><button class="flex items-center gap-1.5 pb-3 text-sm font-medium whitespace-nowrap border-b-2 transition-colors border-transparent text-slate-500 hover:text-slate-700">Agent Skills<span class="text-xs px-1.5 py-0.5 rounded-full bg-slate-100 text-slate-500">24</span></button><button class="flex items-center gap-1.5 pb-3 text-sm font-medium whitespace-nowrap border-b-2 transition-colors border-transparent text-slate-500 hover:text-slate-700">Learn<span class="text-xs px-1.5 py-0.5 rounded-full bg-slate-100 text-slate-500">7</span></button><button class="flex items-center gap-1.5 pb-3 text-sm font-medium whitespace-nowrap border-b-2 transition-colors border-transparent text-slate-500 hover:text-slate-700">PE<span class="text-xs px-1.5 py-0.5 rounded-full bg-slate-100 text-slate-500">5</span></button><button class="flex items-center gap-1.5 pb-3 text-sm font-medium whitespace-nowrap border-b-2 transition-colors border-transparent text-slate-500 hover:text-slate-700">Newsroom<span class="text-xs px-1.5 py-0.5 rounded-full bg-slate-100 text-slate-500">2</span></button><button class="flex items-center gap-1.5 pb-3 text-sm font-medium whitespace-nowrap border-b-2 transition-colors border-transparent text-slate-500 hover:text-slate-700">Legal<span class="text-xs px-1.5 py-0.5 rounded-full bg-slate-100 text-slate-500">1</span></button></div></div></section><section class="max-w-[1200px] mx-auto px-6 py-12"><a class="group block" href="/blog/prior-auth-maze-thirty-rulebooks-six-front-doors"><div class="grid md:grid-cols-2 gap-8 items-center"><div class="relative w-full aspect-[4/3] rounded-xl overflow-hidden"><img alt="The Agentic Practice Issue 7 cover: The Prior Auth Maze You Built Yourself â thirty rulebooks, six front doors, and no map" decoding="async" data-nimg="fill" class="object-cover" style="position:absolute;height:100%;width:100%;left:0;top:0;right:0;bottom:0;color:transparent" sizes="(max-width: 768px) 100vw, 580px" srcSet="/_next/image?url=%2Fimages%2Fblog%2FNews7_image1.png&w=640&q=75 640w, /_next/image?url=%2Fimages%2Fblog%2FNews7_image1.png&w=750&q=75 750w, /_next/image?url=%2Fimages%2Fblog%2FNews7_image1.png&w=828&q=75 828w, /_next/image?url=%2Fimages%2Fblog%2FNews7_image1.png&w=1080&q=75 1080w, /_next/image?url=%2Fimages%2Fblog%2FNews7_image1.png&w=1200&q=75 1200w, /_next/image?url=%2Fimages%2Fblog%2FNews7_image1.png&w=1920&q=75 1920w, /_next/image?url=%2Fimages%2Fblog%2FNews7_image1.png&w=2048&q=75 2048w, /_next/image?url=%2Fimages%2Fblog%2FNews7_image1.png&w=3840&q=75 3840w" src="/_next/image?url=%2Fimages%2Fblog%2FNews7_image1.png&w=3840&q=75"/></div><div><span class="inline-block text-xs font-medium text-agentman-600 mb-3">Healthcare</span><h2 class="text-2xl sm:text-3xl font-bold text-charcoal-950 tracking-tight group-hover:text-agentman-600 transition-colors">Newsletter #7: The Prior Auth Maze You Built Yourself: Thirty Rulebooks, Six Front Doors, and No Map</h2><p class="mt-3 text-slate-600 leading-relaxed line-clamp-3">One continuous glucose monitor, three different submission channels, depending on the payer. After nine years, our practice had six front doors for prior auth and no map â and most of what makes a PA succeed lives in one coordinator's head. Here's the whiteboard workflow we drew before writing a line of code.</p><div class="flex items-center gap-3 mt-5 text-sm text-slate-500"><span class="font-medium text-charcoal-950">Sachin Gangupantula</span><span aria-hidden="true">·</span><time dateTime="2026-09-02T16:00:00.000Z">Sep 2, 2026</time><span aria-hidden="true">·</span><span>15 min read</span></div><span class="inline-flex items-center gap-1.5 mt-5 text-sm font-medium text-charcoal-950 group-hover:text-agentman-600 transition-colors">Read more<svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-arrow-right w-4 h-4 group-hover:translate-x-0.5 transition-transform duration-200"><path d="M5 12h14"></path><path d="m12 5 7 7-7 7"></path></svg></span></div></div></a><div class="border-b border-slate-200 mt-12"></div></section><section class="max-w-[1200px] mx-auto px-6 py-10"><div class="grid grid-cols-1 sm:grid-cols-2 lg:grid-cols-3 gap-6"><a class="group block h-full" href="/blog/why-practices-wait-prior-authorization"><article class="flex flex-col h-full border border-slate-200 rounded-xl overflow-hidden bg-white hover:shadow-lg hover:-translate-y-0.5 transition-all duration-200"><div class="relative w-full aspect-square"><img alt="Two paths drawn as thin charcoal lines from a single starting point. The upper path runs straight to a tan form outline with several empty fields. The lower path curves through a small terracotta circle before reaching an identical form whose fields are already filled with short charcoal lines â the same destination, reached with different amounts of work." loading="lazy" decoding="async" data-nimg="fill" class="object-cover" style="position:absolute;height:100%;width:100%;left:0;top:0;right:0;bottom:0;
10color:transparent" sizes="(max-width: 640px) 100vw, (max-width: 1024px) 50vw, 370px" srcSet="/_next/image?url=%2Fimages%2Fblog%2Fprior-auth-wait-vs-file-hero.jpg&w=384&q=75 384w, /_next/image?url=%2Fimages%2Fblog%2Fprior-auth-wait-vs-file-hero.jpg&w=640&q=75 640w, /_next/image?url=%2Fimages%2Fblog%2Fprior-auth-wait-vs-file-hero.jpg&w=750&q=75 750w, /_next/image?url=%2Fimages%2Fblog%2Fprior-auth-wait-vs-file-hero.jpg&w=828&q=75 828w, /_next/image?url=%2Fimages%2Fblog%2Fprior-auth-wait-vs-file-hero.jpg&w=1080&q=75 1080w, /_next/image?url=%2Fimages%2Fblog%2Fprior-auth-wait-vs-file-hero.jpg&w=1200&q=75 1200w, /_next/image?url=%2Fimages%2Fblog%2Fprior-auth-wait-vs-file-hero.jpg&w=1920&q=75 1920w, /_next/image?url=%2Fimages%2Fblog%2Fprior-auth-wait-vs-file-hero.jpg&w=2048&q=75 2048w, /_next/image?url=%2Fimages%2Fblog%2Fprior-auth-wait-vs-file-hero.jpg&w=3840&q=75 3840w" src="/_next/image?url=%2Fimages%2Fblog%2Fprior-auth-wait-vs-file-hero.jpg&w=3840&q=75"/></div><div class="flex flex-col flex-1 p-5"><div class="flex items-center gap-2 text-xs text-slate-500"><span class="font-medium text-charcoal-950">Prasad Thammineni</span><span aria-hidden="true">·</span><time dateTime="2026-08-10T15:00:00.000Z">Aug 10, 2026</time></div><h3 class="text-[16px] font-semibold leading-snug text-charcoal-950 mt-1.5 group-hover:text-agentman-600 transition-colors line-clamp-2">Should You File a Prior Auth Before the Pharmacy Asks? 83% of Prescribers Say No.</h3><p class="text-sm text-slate-600 mt-2 line-clamp-2 flex-1">CoverMyMeds publishes that patients get their medications 13.2 days sooner when prior auths are started at the point of prescribing. Their own survey found only 17% of providers do it. The gap isn't ignorance â there's a real argument for waiting that nobody has written down.</p><div class="flex items-center justify-between mt-4 pt-3 border-t border-slate-100"><span class="text-xs font-medium text-agentman-600">Healthcare</span><svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-arrow-right w-4 h-4 text-slate-400 group-hover:text-agentman-600 group-hover:translate-x-0.5 transition-all duration-200"><path d="M5 12h14"></path><path d="m12 5 7 7-7 7"></path></svg></div></div></article></a><a class="group block h-full" href="/blog/refill-and-prior-auth-same-fax"><article class="flex flex-col h-full border border-slate-200 rounded-xl overflow-hidden bg-white hover:shadow-lg hover:-translate-y-0.5 transition-all duration-200"><div class="relative w-full aspect-square"><img alt="Two tan documents with folded corners sit side by side, joined beneath by a single thin charcoal line that loops from one to the other and back, with a small terracotta circle at the loop's midpoint â two separate pages describing one underlying event." loading="lazy" decoding="async" data-nimg="fill" class="object-cover" style="position:absolute;height:100%;width:100%;left:0;top:0;right:0;bottom:0;color:transparent" sizes="(max-width: 640px) 100vw, (max-width: 1024px) 50vw, 370px" srcSet="/_next/image?url=%2Fimages%2Fblog%2Frefill-pa-same-fax-hero.jpg&w=384&q=75 384w, /_next/image?url=%2Fimages%2Fblog%2Frefill-pa-same-fax-hero.jpg&w=640&q=75 640w, /_next/image?url=%2Fimages%2Fblog%2Frefill-pa-same-fax-hero.jpg&w=750&q=75 750w, /_next/image?url=%2Fimages%2Fblog%2Frefill-pa-same-fax-hero.jpg&w=828&q=75 828w, /_next/image?url=%2Fimages%2Fblog%2Frefill-pa-same-fax-hero.jpg&w=1080&q=75 1080w, /_next/image?url=%2Fimages%2Fblog%2Frefill-pa-same-fax-hero.jpg&w=1200&q=75 1200w, /_next/image?url=%2Fimages%2Fblog%2Frefill-pa-same-fax-hero.jpg&w=1920&q=75 1920w, /_next/image?url=%2Fimages%2Fblog%2Frefill-pa-same-fax-hero.jpg&w=2048&q=75 2048w, /_next/image?url=%2Fimages%2Fblog%2Frefill-pa-same-fax-hero.jpg&w=3840&q=75 3840w" src="/_next/image?url=%2Fimages%2Fblog%2Frefill-pa-same-fax-hero.jpg&w=3840&q=75"/></div><div class="flex flex-col flex-1 p-5"><div class="flex items-center gap-2 text-xs text-slate-500"><span class="font-medium text-charcoal-950">Prasad Thammineni</span><span aria-hidden="true">·</span><time dateTime="2026-08-10T14:00:00.000Z">Aug 10, 2026</time></div><h3 class="text-[16px] font-semibold leading-snug text-charcoal-950 mt-1.5 group-hover:text-agentman-600 transition-colors line-clamp-2">
10The Refill and the Prior Auth Are the Same Fax, Arriving Twice</h3><p class="text-sm text-slate-600 mt-2 line-clamp-2 flex-1">A prescription goes out. The pharmacy tries to fill it. The claim rejects. That rejection generates a prior authorization request that arrives back at the practice as a fax â the same clinical event, now a second document, landing in the same pile as everything else.</p><div class="flex items-center justify-between mt-4 pt-3 border-t border-slate-100"><span class="text-xs font-medium text-agentman-600">Healthcare</span><svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-arrow-right w-4 h-4 text-slate-400 group-hover:text-agentman-600 group-hover:translate-x-0.5 transition-all duration-200"><path d="M5 12h14"></path><path d="m12 5 7 7-7 7"></path></svg></div></div></article></a><a class="group block h-full" href="/blog/third-of-refill-requests-are-not-refills"><article class="flex flex-col h-full border border-slate-200 rounded-xl overflow-hidden bg-white hover:shadow-lg hover:-translate-y-0.5 transition-all duration-200"><div class="relative w-full aspect-square"><img alt="Ten small outlined pill-shaped tiles arranged in a grid on cream; six are plain tan, and four are terracotta, each marked with a small charcoal symbol â a calendar, a test tube, an upward arrow, and a cross â showing that a substantial share of routine-looking requests carry a finding." loading="lazy" decoding="async" data-nimg="fill" class="object-cover" style="position:absolute;height:100%;width:100%;left:0;top:0;right:0;bottom:0;color:transparent" sizes="(max-width: 640px) 100vw, (max-width: 1024px) 50vw, 370px" srcSet="/_next/image?url=%2Fimages%2Fblog%2Fthird-of-refills-clinical-hero.jpg&w=384&q=75 384w, /_next/image?url=%2Fimages%2Fblog%2Fthird-of-refills-clinical-hero.jpg&w=640&q=75 640w, /_next/image?url=%2Fimages%2Fblog%2Fthird-of-refills-clinical-hero.jpg&w=750&q=75 750w, /_next/image?url=%2Fimages%2Fblog%2Fthird-of-refills-clinical-hero.jpg&w=828&q=75 828w, /_next/image?url=%2Fimages%2Fblog%2Fthird-of-refills-clinical-hero.jpg&w=1080&q=75 1080w, /_next/image?url=%2Fimages%2Fblog%2Fthird-of-refills-clinical-hero.jpg&w=1200&q=75 1200w, /_next/image?url=%2Fimages%2Fblog%2Fthird-of-refills-clinical-hero.jpg&w=1920&q=75 1920w, /_next/image?url=%2Fimages%2Fblog%2Fthird-of-refills-clinical-hero.jpg&w=2048&q=75 2048w, /_next/image?url=%2Fimages%2Fblog%2Fthird-of-refills-clinical-hero.jpg&w=3840&q=75 3840w" src="/_next/image?url=%2Fimages%2Fblog%2Fthird-of-refills-clinical-hero.jpg&w=3840&q=75"/></div><div class="flex flex-col flex-1 p-5"><div class="flex items-center gap-2 text-xs text-slate-500"><span class="font-medium text-charcoal-950">Prasad Thammineni</span><span aria-hidden="true">·</span><time dateTime="2026-08-09T15:00:00.000Z">Aug 9, 2026</time></div><h3 class="text-[16px] font-semibold leading-snug text-charcoal-950 mt-1.5 group-hover:text-agentman-600 transition-colors line-clamp-2">A Third of Refill Requests Aren't Refill Requests</h3><p class="text-sm text-slate-600 mt-2 line-clamp-2 flex-1">In a year-long study at a federally qualified health center, 36% of refill requests carried a clinical finding â a patient overdue for labs, a dose that should have changed, a drug that should have stopped. The more troubling number: of the drug therapy problems found, 55% were never addressed.</p><div class="flex items-center justify-between mt-4 pt-3 border-t border-slate-100"><span class="text-xs font-medium text-agentman-600">Healthcare</span><svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-arrow-right w-4 h-4 text-slate-400 group-hover:text-agentman-600 group-hover:translate-x-0.5 transition-all duration-200"><path d="M5 12h14"></path><path d="m12 5 7 7-7 7"></path></svg></div></div></article></a><a class="group block h-full" href="/blog/refill-request-is-not-a-task"><article class="flex flex-col h-full border border-slate-200 rounded-xl overflow-hidden bg-white hover:shadow-lg hover:-translate-y-0.5 transition-all duration-200"><div class="relative w-full aspect-square"><img alt="A single tan document at the left joined by thin charcoal lines that fan out to six small outlined shapes: a checkmark, a calendar, a vial, a clipboard form, two mismatched rectangles, and an arrow turning away â one arriving request resolving into six different kinds of work." loading="lazy" decoding="async" data-nimg="fill" class="object-cover" style="position:absolute;height:100%;width:100%;left:0;top:0;right:0;bottom:0;
10color:transparent" sizes="(max-width: 640px) 100vw, (max-width: 1024px) 50vw, 370px" srcSet="/_next/image?url=%2Fimages%2Fblog%2Frefill-unclassified-trigger-hero.jpg&w=384&q=75 384w, /_next/image?url=%2Fimages%2Fblog%2Frefill-unclassified-trigger-hero.jpg&w=640&q=75 640w, /_next/image?url=%2Fimages%2Fblog%2Frefill-unclassified-trigger-hero.jpg&w=750&q=75 750w, /_next/image?url=%2Fimages%2Fblog%2Frefill-unclassified-trigger-hero.jpg&w=828&q=75 828w, /_next/image?url=%2Fimages%2Fblog%2Frefill-unclassified-trigger-hero.jpg&w=1080&q=75 1080w, /_next/image?url=%2Fimages%2Fblog%2Frefill-unclassified-trigger-hero.jpg&w=1200&q=75 1200w, /_next/image?url=%2Fimages%2Fblog%2Frefill-unclassified-trigger-hero.jpg&w=1920&q=75 1920w, /_next/image?url=%2Fimages%2Fblog%2Frefill-unclassified-trigger-hero.jpg&w=2048&q=75 2048w, /_next/image?url=%2Fimages%2Fblog%2Frefill-unclassified-trigger-hero.jpg&w=3840&q=75 3840w" src="/_next/image?url=%2Fimages%2Fblog%2Frefill-unclassified-trigger-hero.jpg&w=3840&q=75"/></div><div class="flex flex-col flex-1 p-5"><div class="flex items-center gap-2 text-xs text-slate-500"><span class="font-medium text-charcoal-950">Prasad Thammineni</span><span aria-hidden="true">·</span><time dateTime="2026-08-09T14:00:00.000Z">Aug 9, 2026</time></div><h3 class="text-[16px] font-semibold leading-snug text-charcoal-950 mt-1.5 group-hover:text-agentman-600 transition-colors line-clamp-2">A Refill Request Isn't a Task. It's an Unclassified Trigger.</h3><p class="text-sm text-slate-600 mt-2 line-clamp-2 flex-1">Every vendor selling refill automation answers the same question: what percentage can you auto-approve? It's the wrong question. A refill request has no identity until someone reads it â and in our production triage system, it can turn out to be any of six different jobs.</p><div class="flex items-center justify-between mt-4 pt-3 border-t border-slate-100"><span class="text-xs font-medium text-agentman-600">Healthcare</span><svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-arrow-right w-4 h-4 text-slate-400 group-hover:text-agentman-600 group-hover:translate-x-0.5 transition-all duration-200"><path d="M5 12h14"></path><path d="m12 5 7 7-7 7"></path></svg></div></div></article></a><a class="group block h-full" href="/blog/denial-report-cannot-see-denials"><article class="flex flex-col h-full border border-slate-200 rounded-xl overflow-hidden bg-white hover:shadow-lg hover:-translate-y-0.5 transition-all duration-200"><div class="relative w-full aspect-square"><img alt="Six bars on a baseline inside a thin rectangular report frame. Five are short and tan; one, in terracotta, is several times taller and breaks straight through the top edge of the frame â the denials a payment-derived report cannot contain or show." loading="lazy" decoding="async" data-nimg="fill" class="object-cover" style="position:absolute;height:100%;width:100%;left:0;top:0;right:0;bottom:0;color:transparent" sizes="(max-width: 640px) 100vw, (max-width: 1024px) 50vw, 370px" srcSet="/_next/image?url=%2Fimages%2Fblog%2Fdenial-report-blind-hero.jpg&w=384&q=75 384w, /_next/image?url=%2Fimages%2Fblog%2Fdenial-report-blind-hero.jpg&w=640&q=75 640w, /_next/image?url=%2Fimages%2Fblog%2Fdenial-report-blind-hero.jpg&w=750&q=75 750w, /_next/image?url=%2Fimages%2Fblog%2Fdenial-report-blind-hero.jpg&w=828&q=75 828w, /_next/image?url=%2Fimages%2Fblog%2Fdenial-report-blind-hero.jpg&w=1080&q=75 1080w, /_next/image?url=%2Fimages%2Fblog%2Fdenial-report-blind-hero.jpg&w=1200&q=75 1200w, /_next/image?url=%2Fimages%2Fblog%2Fdenial-report-blind-hero.jpg&w=1920&q=75 1920w, /_next/image?url=%2Fimages%2Fblog%2Fdenial-report-blind-hero.jpg&w=2048&q=75 2048w, /_next/image?url=%2Fimages%2Fblog%2Fdenial-report-blind-hero.jpg&w=3840&q=75 3840w" src="/_next/image?url=%2Fimages%2Fblog%2Fdenial-report-blind-hero.jpg&w=3840&q=75"/></div><div class="flex flex-col flex-1 p-5"><div class="flex items-center gap-2 text-xs text-slate-500"><span class="font-medium text-charcoal-950">Prasad Thammineni</span><span aria-hidden="true">·</span><time dateTime="2026-08-08T14:00:00.000Z">Aug 8, 2026</time></div><h3 class="text-[16px] font-semibold leading-snug text-charcoal-950 mt-1.5 group-hover:text-agentman-600 transition-colors line-clamp-2">Your Denial Report Can't See Denials. Here's the Proof.</h3><p class="text-sm text-slate-600 mt-2 line-clamp-2 flex-1">We analyzed one practice's denials and found $98 worth worth appealing. Then we read the actual remittances and found $5,413 â in the same month, at the same practice. The first method was structurally blind, and it's the method most denial reports use.</p><div class="flex items-center justify-between mt-4 pt-3 border-t border-slate-100"><span class="text-xs font-medium text-agentman-600">Healthcare</span><svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-arrow-right w-4 h-4 text-slate-400 group-hover:text-agentman-600 group-hover:translate-x-0.5 transition-all duration-200"><path d="M5 12h14"></path><path d="m12 5 7 7-7 7"></path></svg></div></div></article></a><a class="group block h-full" href="/blog/security-questions-before-signing-ai-vendor"><article class="flex flex-col h-full border border-slate-200 rounded-xl overflow-hidden bg-white hover:shadow-lg hover:-translate-y-0.5 transition-all duration-200"><div class="relative w-full aspect-square"><img alt="A tall tan checklist document outlined in charcoal with a column of short horizontal lines beside small square checkboxes, three of them marked with terracotta checks, and a small terracotta magnifying glass resting over the lower portion of the list." loading="lazy" decoding="async" data-nimg="fill" class="object-cover" style="position:absolute;height:100%;width:100%;left:0;top:0;right:0;bottom:0;
10color:transparent" sizes="(max-width: 640px) 100vw, (max-width: 1024px) 50vw, 370px" srcSet="/_next/image?url=%2Fimages%2Fblog%2Fsecurity-questions-ai-vendor-hero.jpg&w=384&q=75 384w, /_next/image?url=%2Fimages%2Fblog%2Fsecurity-questions-ai-vendor-hero.jpg&w=640&q=75 640w, /_next/image?url=%2Fimages%2Fblog%2Fsecurity-questions-ai-vendor-hero.jpg&w=750&q=75 750w, /_next/image?url=%2Fimages%2Fblog%2Fsecurity-questions-ai-vendor-hero.jpg&w=828&q=75 828w, /_next/image?url=%2Fimages%2Fblog%2Fsecurity-questions-ai-vendor-hero.jpg&w=1080&q=75 1080w, /_next/image?url=%2Fimages%2Fblog%2Fsecurity-questions-ai-vendor-hero.jpg&w=1200&q=75 1200w, /_next/image?url=%2Fimages%2Fblog%2Fsecurity-questions-ai-vendor-hero.jpg&w=1920&q=75 1920w, /_next/image?url=%2Fimages%2Fblog%2Fsecurity-questions-ai-vendor-hero.jpg&w=2048&q=75 2048w, /_next/image?url=%2Fimages%2Fblog%2Fsecurity-questions-ai-vendor-hero.jpg&w=3840&q=75 3840w" src="/_next/image?url=%2Fimages%2Fblog%2Fsecurity-questions-ai-vendor-hero.jpg&w=3840&q=75"/></div><div class="flex flex-col flex-1 p-5"><div class="flex items-center gap-2 text-xs text-slate-500"><span class="font-medium text-charcoal-950">Prasad Thammineni</span><span aria-hidden="true">·</span><time dateTime="2026-08-08T00:00:00.000Z">Aug 8, 2026</time></div><h3 class="text-[16px] font-semibold leading-snug text-charcoal-950 mt-1.5 group-hover:text-agentman-600 transition-colors line-clamp-2">21 Questions to Ask an AI Vendor Before You Sign</h3><p class="text-sm text-slate-600 mt-2 line-clamp-2 flex-1">A security questionnaire a small practice can actually run â no procurement team required. Copy the list, send it, and read the answers for artifacts rather than adjectives. We answer all 21 ourselves at the bottom.</p><div class="flex items-center justify-between mt-4 pt-3 border-t border-slate-100"><span class="text-xs font-medium text-agentman-600">Healthcare</span><svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-arrow-right w-4 h-4 text-slate-400 group-hover:text-agentman-600 group-hover:translate-x-0.5 transition-all duration-200"><path d="M5 12h14"></path><path d="m12 5 7 7-7 7"></path></svg></div></div></article></a><a class="group block h-full" href="/blog/can-ai-read-phi-data-handling-rules"><article class="flex flex-col h-full border border-slate-200 rounded-xl overflow-hidden bg-white hover:shadow-lg hover:-translate-y-0.5 transition-all duration-200"><div class="relative w-full aspect-square"><img alt="A tall tan document with several lines of text, most of them covered by solid charcoal redaction bars, and a single terracotta-outlined magnifying glass hovering over the two lines left visible â an agent seeing only the minimum necessary fields." loading="lazy" decoding="async" data-nimg="fill" class="object-cover" style="position:absolute;height:100%;width:100%;left:0;top:0;right:0;bottom:0;color:transparent" sizes="(max-width: 640px) 100vw, (max-width: 1024px) 50vw, 370px" srcSet="/_next/image?url=%2Fimages%2Fblog%2Fcan-ai-read-phi-hero.jpg&w=384&q=75 384w, /_next/image?url=%2Fimages%2Fblog%2Fcan-ai-read-phi-hero.jpg&w=640&q=75 640w, /_next/image?url=%2Fimages%2Fblog%2Fcan-ai-read-phi-hero.jpg&w=750&q=75 750w, /_next/image?url=%2Fimages%2Fblog%2Fcan-ai-read-phi-hero.jpg&w=828&q=75 828w, /_next/image?url=%2Fimages%2Fblog%2Fcan-ai-read-phi-hero.jpg&w=1080&q=75 1080w, /_next/image?url=%2Fimages%2Fblog%2Fcan-ai-read-phi-hero.jpg&w=1200&q=75 1200w, /_next/image?url=%2Fimages%2Fblog%2Fcan-ai-read-phi-hero.jpg&w=1920&q=75 1920w, /_next/image?url=%2Fimages%2Fblog%2Fcan-ai-read-phi-hero.jpg&w=2048&q=75 2048w, /_next/image?url=%2Fimages%2Fblog%2Fcan-ai-read-phi-hero.jpg&w=3840&q=75 3840w" src="/_next/image?url=%2Fimages%2Fblog%2Fcan-ai-read-phi-hero.jpg&w=3840&q=75"/></div><div class="flex flex-col flex-1 p-5"><div class="flex items-center gap-2 text-xs text-slate-500"><span class="font-medium text-charcoal-950">Prasad Thammineni</span><span aria-hidden="true">·</span><time dateTime="2026-08-07T23:30:00.000Z">Aug 7, 2026</time></div><h3 class="text-[16px] font-semibold leading-snug text-charcoal-950 mt-1.5 group-hover:text-agentman-600 transition-colors line-clamp-2">Can AI Read PHI? The Data-Handling Rules That Actually Apply</h3><p class="text-sm text-slate-600 mt-2 line-clamp-2 flex-1">Yes â with a BAA and the right safeguards. The harder questions are where the data travels during inference, how little of it the agent actually needs, and what ends up sitting in prompt logs nobody thought to check.</p><div class="flex items-center justify-between mt-4 pt-3 border-t border-slate-100"><span class="text-xs font-medium text-agentman-600">Healthcare</span><svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-arrow-right w-4 h-4 text-slate-400 group-hover:text-agentman-600 group-hover:translate-x-0.5 transition-all duration-200"><path d="M5 12h14"></path><path d="m12 5 7 7-7 7"></path></svg></div></div></article></a><a class="group block h-full" href="/blog/soc-2-for-ai-vendors"><article class="flex flex-col h-full border border-slate-200 rounded-xl overflow-hidden bg-white hover:shadow-lg hover:-translate-y-0.5 transition-all duration-200"><div class="relative w-full aspect-square"><img alt="A tan badge shape on the left with a small terracotta question mark inside it, and on the right a multi-page report document with a visible date range bar and several short lines, connected by a thin charcoal arrow pointing from badge to report." loading="lazy" decoding="async" data-nimg="fill" class="object-cover" style="position:absolute;height:100%;width:100%;left:0;top:0;right:0;bottom:0;
10color:transparent" sizes="(max-width: 640px) 100vw, (max-width: 1024px) 50vw, 370px" srcSet="/_next/image?url=%2Fimages%2Fblog%2Fsoc-2-for-ai-vendors-hero.jpg&w=384&q=75 384w, /_next/image?url=%2Fimages%2Fblog%2Fsoc-2-for-ai-vendors-hero.jpg&w=640&q=75 640w, /_next/image?url=%2Fimages%2Fblog%2Fsoc-2-for-ai-vendors-hero.jpg&w=750&q=75 750w, /_next/image?url=%2Fimages%2Fblog%2Fsoc-2-for-ai-vendors-hero.jpg&w=828&q=75 828w, /_next/image?url=%2Fimages%2Fblog%2Fsoc-2-for-ai-vendors-hero.jpg&w=1080&q=75 1080w, /_next/image?url=%2Fimages%2Fblog%2Fsoc-2-for-ai-vendors-hero.jpg&w=1200&q=75 1200w, /_next/image?url=%2Fimages%2Fblog%2Fsoc-2-for-ai-vendors-hero.jpg&w=1920&q=75 1920w, /_next/image?url=%2Fimages%2Fblog%2Fsoc-2-for-ai-vendors-hero.jpg&w=2048&q=75 2048w, /_next/image?url=%2Fimages%2Fblog%2Fsoc-2-for-ai-vendors-hero.jpg&w=3840&q=75 3840w" src="/_next/image?url=%2Fimages%2Fblog%2Fsoc-2-for-ai-vendors-hero.jpg&w=3840&q=75"/></div><div class="flex flex-col flex-1 p-5"><div class="flex items-center gap-2 text-xs text-slate-500"><span class="font-medium text-charcoal-950">Prasad Thammineni</span><span aria-hidden="true">·</span><time dateTime="2026-08-07T23:00:00.000Z">Aug 7, 2026</time></div><h3 class="text-[16px] font-semibold leading-snug text-charcoal-950 mt-1.5 group-hover:text-agentman-600 transition-colors line-clamp-2">SOC 2 for AI Vendors: What It Covers and What It Doesn't</h3><p class="text-sm text-slate-600 mt-2 line-clamp-2 flex-1">A SOC 2 report is real evidence â but the vendor helped choose what was examined, and the badge on the website tells you none of that. What to ask for, how to read the report, and why 'SOC 2 ready' means nothing at all.</p><div class="flex items-center justify-between mt-4 pt-3 border-t border-slate-100"><span class="text-xs font-medium text-agentman-600">Healthcare</span><svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-arrow-right w-4 h-4 text-slate-400 group-hover:text-agentman-600 group-hover:translate-x-0.5 transition-all duration-200"><path d="M5 12h14"></path><path d="m12 5 7 7-7 7"></path></svg></div></div></article></a><a class="group block h-full" href="/blog/hipaa-compliant-ai-agents-baa-phi-audit-trails"><article class="flex flex-col h-full border border-slate-200 rounded-xl overflow-hidden bg-white hover:shadow-lg hover:-translate-y-0.5 transition-all duration-200"><div class="relative w-full aspect-square"><img alt="Three elements in a row joined by thin lines: a tan document with signature marks, a horizontal chain of three linked tiles, and a block of stacked terracotta dashes beside a vertical rule â the signed agreement, the subprocessor chain it binds, and the audit log that records what happened." loading="lazy" decoding="async" data-nimg="fill" class="object-cover" style="position:absolute;height:100%;width:100%;left:0;top:0;right:0;bottom:0;color:transparent" sizes="(max-width: 640px) 100vw, (max-width: 1024px) 50vw, 370px" srcSet="/_next/image?url=%2Fimages%2Fblog%2Fhipaa-ai-agents-baa-audit-hero.jpg&w=384&q=75 384w, /_next/image?url=%2Fimages%2Fblog%2Fhipaa-ai-agents-baa-audit-hero.jpg&w=640&q=75 640w, /_next/image?url=%2Fimages%2Fblog%2Fhipaa-ai-agents-baa-audit-hero.jpg&w=750&q=75 750w, /_next/image?url=%2Fimages%2Fblog%2Fhipaa-ai-agents-baa-audit-hero.jpg&w=828&q=75 828w, /_next/image?url=%2Fimages%2Fblog%2Fhipaa-ai-agents-baa-audit-hero.jpg&w=1080&q=75 1080w, /_next/image?url=%2Fimages%2Fblog%2Fhipaa-ai-agents-baa-audit-hero.jpg&w=1200&q=75 1200w, /_next/image?url=%2Fimages%2Fblog%2Fhipaa-ai-agents-baa-audit-hero.jpg&w=1920&q=75 1920w, /_next/image?url=%2Fimages%2Fblog%2Fhipaa-ai-agents-baa-audit-hero.jpg&w=2048&q=75 2048w, /_next/image?url=%2Fimages%2Fblog%2Fhipaa-ai-agents-baa-audit-hero.jpg&w=3840&
10q=75 3840w" src="/_next/image?url=%2Fimages%2Fblog%2Fhipaa-ai-agents-baa-audit-hero.jpg&w=3840&q=75"/></div><div class="flex flex-col flex-1 p-5"><div class="flex items-center gap-2 text-xs text-slate-500"><span class="font-medium text-charcoal-950">Prasad Thammineni</span><span aria-hidden="true">·</span><time dateTime="2026-08-07T22:30:00.000Z">Aug 7, 2026</time></div><h3 class="text-[16px] font-semibold leading-snug text-charcoal-950 mt-1.5 group-hover:text-agentman-600 transition-colors line-clamp-2">HIPAA-Compliant AI Agents: BAAs, PHI Handling, and Audit Trails</h3><p class="text-sm text-slate-600 mt-2 line-clamp-2 flex-1">An AI agent that touches patient data is a business associate, with direct legal liability. Here's what the BAA has to cover, which AI-specific questions older templates miss, and what an audit trail needs to contain when an agent acted instead of a person.</p><div class="flex items-center justify-between mt-4 pt-3 border-t border-slate-100"><span class="text-xs font-medium text-agentman-600">Healthcare</span><svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-arrow-right w-4 h-4 text-slate-400 group-hover:text-agentman-600 group-hover:translate-x-0.5 transition-all duration-200"><path d="M5 12h14"></path><path d="m12 5 7 7-7 7"></path></svg></div></div></article></a></div></section><section class="max-w-[1200px] mx-auto px-6 pb-16"><nav aria-label="Blog pagination" class="flex items-center justify-center gap-2 mt-12"><button disabled="" aria-label="Previous page" class="inline-flex items-center justify-center w-9 h-9 rounded-lg border border-slate-200 text-slate-600 hover:bg-slate-50 disabled:opacity-40 disabled:cursor-not-allowed transition-colors"><svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-chevron-left w-4 h-4"><path d="m15 18-6-6 6-6"></path></svg></button><button aria-current="page" class="inline-flex items-center justify-center w-9 h-9 rounded-lg text-sm font-medium transition-colors bg-charcoal-950 text-white">1</button><button class="inline-flex items-center justify-center w-9 h-9 rounded-lg text-sm font-medium transition-colors text-slate-600 hover:bg-slate-100">2</button><span class="px-1 text-slate-400">â¦</span><button class="inline-flex items-center justify-center w-9 h-9 rounded-lg text-sm font-medium transition-colors text-slate-600 hover:bg-slate-100">10</button><button aria-label="Next page" class="inline-flex items-center justify-center w-9 h-9 rounded-lg border border-slate-200 text-slate-600 hover:bg-slate-50 disabled:opacity-40 disabled:cursor-not-allowed transition-colors"><svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-chevron-right w-4 h-4"><path d="m9 18 6-6-6-6"></path></svg></button><span class="ml-3 text-sm text-slate-500">of <!-- -->10</span></nav></section></main><section id="newsletter" class="bg-agentman-75 py-16 scroll-mt-24"><div class="max-w-2xl mx-auto px-6 text-center"><h2 class="text-2xl sm:text-3xl font-bold text-charcoal-950 tracking-tight mb-3">Stay in the loop</h2><p class="text-slate-600 mb-8">Get practical insights on production AI agents delivered to your inbox.</p><form class="flex flex-col sm:flex-row gap-3 max-w-md mx-auto"><input type="email" required="" placeholder="[email protected]" class="flex-1 px-4 py-2.5 rounded-lg border border-slate-300 text-sm focus:outline-none focus:ring-2 focus:ring-agentman-500 focus:border-transparent" value=""/><button type="submit" class="px-5 py-2.5 rounded-lg bg-charcoal-950 text-white text-sm font-medium hover:bg-charcoal-900 disabled:opacity-60 transition-colors">Subscribe</button></form></div></section><section class="bg-charcoal-950 py-16 text-center"><div class="max-w-2xl mx-auto px-6"><h2 class="text-2xl sm:text-3xl font-bold text-white tracking-tight mb-4">Ready to automate your back office?</h2><p class="text-white/60 text-lg mb-8">See how production-grade AI agents handle your toughest workflows.</p><div class="flex flex-col sm:flex-row items-center justify-center gap-4"><a href="mailto:[email protected]" class="inline-flex items-center justify-center px-6 py-3 rounded-lg bg-agentman-500 hover:bg-agentman-600 text-white font-medium transition-colors">Talk to us</a><a class="inline-flex items-center justify-center px-6 py-3 rounded-lg border border-white/20 text-white hover:bg-white/10 font-medium transition-colors" href="/platform">See the platform</a></div></div></section><footer class="relative bg-white border-t border-gray-200"><div class="relative max-w-7xl mx-auto py-16 px-4 sm:px-6 lg:px-8"><div class="grid grid-cols-1 md:grid-cols-2 lg:grid-cols-6 gap-8 mb-12"><div class="md:col-span-1 lg:pr-8"><a class="block mb-4" href="/"><img alt="Agentman" loading="lazy" width="160" height="37" decoding="async" data-nimg="1" class="h-9 w-auto" style="color:transparent;object-fit:contain" srcSet="/_next/image?url=%2Fagentman-logo-mark-simplified.svg&w=256&q=75 1x, /_next/image?url=%2Fagentman-logo-mark-simplified.svg&w=384&q=75 2x" src="/_next/image?url=%2Fagentman-logo-mark-simplified.svg&w=384&q=75"/></a><p class="text-sm text-charcoal-950 mb-6">Bring joy back to your practice.</p><div class="flex flex-col gap-3"><div class="flex gap-3"><a aria-label="Linkedin" target="_blank" rel="noopener noreferrer" class="w-10 h-10 rounded-lg bg-white border border-gray-200 hover:border-agentman-600 flex items-center justify-center transition-all duration-200" href="https://www.linkedin.com/company/agentman/"><svg fill="currentColor" viewBox="0 0 24 24" class="w-5 h-5 text-charcoal-950 hover:text-agentman-600 transition-colors duration-200"><path d="M19 0h-14c-2.761 0-5 2.239-5 5v14c0 2.761 2.239 5 5 5h14c2.762 0 5-2.239 5-5v-14c0-2.761-2.238-5-5-5zm-11 19h-3v-11h3v11zm-1.5-12.268c-.966 0-1.75-.79-1.75-1.764s.784-1.764 1.75-1.764 1.75.79 1.75 1.764-.783 1.764-1.75 1.764zm13.5 12.268h-3v-5.604c0-3.368-4-3.113-4 0v5.604h-3v-11h3v1.765c1.396-2.586 7-2.777 7 2.476v6.759z"></path></svg></a><a aria-label="YouTube" target="_blank" rel="noopener noreferrer" class="w-10 h-10 rounded-lg bg-white border border-gray-200 hover:border-agentman-600 flex items-center justify-center transition-all duration-200" href="https://www.youtube.com/@agentman-ai"><svg fill="currentColor" viewBox="0 0 24 24" class="w-5 h-5 text-charcoal-950 hover:text-agentman-600 transition-colors duration-200"><path d="M23.498 6.186a3.016 3.016 0 0 0-2.122-2.136C19.505 3.545 12 3.545 12 3.545s-7.505 0-9.377.505A3.017 3.017 0 0 0 .502 6.186C0 8.07 0 12 0 12s0 3.93.502 5.814a3.016 3.016 0 0 0 2.122 2.136c1.871.505 9.376.505 9.376.505s7.505 0 9.377-.505a3.015 3.015 0 0 0 2.122-2.136C24 15.93 24 12 24 12s0-3.93-.502-5.814zM9.545 15.568V8.432L15.818 12l-6.273 3.568z"></path></svg></a><a aria-label="Facebook" target="_blank" rel="noopener noreferrer" class="w-10 h-10 rounded-lg bg-white border border-gray-200 hover:border-agentman-600 flex items-center justify-center transition-all duration-200" href="https://www.facebook.com/profile.php?id=61591020467195"><svg fill="currentColor" viewBox="0 0 24 24" class="w-5 h-5 text-charcoal-950 hover:text-agentman-600 transition-colors duration-200"><path d="M24 12.073c0-6.627-5.373-12-12-12s-12 5.373-12 12c0 5.99 4.388 10.954 10.125 11.854v-8.385H7.078v-3.47h3.047V9.43c0-3.007 1.792-4.669 4.533-4.669 1.312 0 2.686.235 2.686.235v2.953H15.83c-1.491 0-1.956.925-1.956 1.874v2.25h3.328l-.532 3.47h-2.796v8.385C19.612 23.027 24 18.062 24 12.073z"></path></svg></a></div><div class="flex gap-3"><a aria-label="TikTok" target="_blank" rel="noopener noreferrer" class="w-10 h-10 rounded-lg bg-white border border-gray-200 hover:border-agentman-600 flex items-center justify-center transition-all duration-200" href="https://www.tiktok.com/@agentman_ai"><svg fill="currentColor" viewBox="0 0 24 24" class="w-5 h-5 text-charcoal-950 hover:text-agentman-600 transition-colors duration-200"><path d="M12.525.02c1.31-.02 2.61-.01 3.91-.02.08 1.53.63 3.09 1.75 4.17 1.12 1.11 2.7 1.62 4.24 1.79v4.03c-1.44-.05-2.89-.35-4.2-.97-.57-.26-1.1-.59-1.62-.93-.01 2.92.01 5.84-.02 8.75-.08 1.4-.54 2.79-1.35 3.94-1.31 1.92-3.58 3.17-5.91 3.21-1.43.08-2.86-.31-4.08
10-1.03-2.02-1.19-3.44-3.37-3.65-5.71-.02-.5-.03-1-.01-1.49.18-1.9 1.12-3.72 2.58-4.96 1.66-1.44 3.98-2.13 6.15-1.72.02 1.48-.04 2.96-.04 4.44-.99-.32-2.15-.23-3.02.37-.63.41-1.11 1.04-1.36 1.75-.21.51-.15 1.07-.14 1.61.24 1.64 1.82 3.02 3.5 2.87 1.12-.01 2.19-.66 2.77-1.61.19-.33.4-.67.41-1.06.1-1.79.06-3.57.07-5.36.01-4.03-.01-8.05.02-12.07z"></path></svg></a><a aria-label="Twitter" target="_blank" rel="noopener noreferrer" class="w-10 h-10 rounded-lg bg-white border border-gray-200 hover:border-agentman-600 flex items-center justify-center transition-all duration-200" href="https://twitter.com/agentman_ai"><svg fill="currentColor" viewBox="0 0 24 24" class="w-5 h-5 text-charcoal-950 hover:text-agentman-600 transition-colors duration-200"><path d="M8.29 20.251c7.547 0 11.675-6.253 11.675-11.675 0-.178 0-.355-.012-.53A8.348 8.348 0 0022 5.92a8.19 8.19 0 01-2.357.646 4.118 4.118 0 001.804-2.27 8.224 8.224 0 01-2.605.996 4.107 4.107 0 00-6.993 3.743 11.65 11.65 0 01-8.457-4.287 4.106 4.106 0 001.27 5.477A4.072 4.072 0 012.8 9.713v.052a4.105 4.105 0 003.292 4.022 4.095 4.095 0 01-1.853.07 4.108 4.108 0 003.834 2.85A8.233 8.233 0 012 18.407a11.616 11.616 0 006.29 1.84"></path></svg></a><a aria-label="Instagram" target="_blank" rel="noopener noreferrer" class="w-10 h-10 rounded-lg bg-white border border-gray-200 hover:border-agentman-600 flex items-center justify-center transition-all duration-200" href="https://www.instagram.com/agentman.ai?utm_source=qr"><svg fill="currentColor" viewBox="0 0 24 24" class="w-5 h-5 text-charcoal-950 hover:text-agentman-600 transition-colors duration-200"><path d="M12 2.163c3.204 0 3.584.012 4.85.07 3.252.148 4.771 1.691 4.919 4.919.058 1.265.069 1.645.069 4.849 0 3.205-.012 3.584-.069 4.849-.149 3.225-1.664 4.771-4.919 4.919-1.266.058-1.644.07-4.85.07-3.204 0-3.584-.012-4.849-.07-3.26-.149-4.771-1.699-4.919-4.92-.058-1.265-.07-1.644-.07-4.849 0-3.204.013-3.583.07-4.849.149-3.227 1.664-4.771 4.919-4.919 1.266-.057 1.645-.069 4.849-.069zm0-2.163c-3.259 0-3.667.014-4.947.072-4.358.2-6.78 2.618-6.98 6.98-.059 1.281-.073 1.689-.073 4.948 0 3.259.014 3.668.072 4.948.2 4.358 2.618 6.78 6.98 6.98 1.281.058 1.689.072 4.948.072 3.259 0 3.668-.014 4.948-.072 4.354-.2 6.782-2.618 6.979-6.98.059-1.28.073-1.689.073-4.948 0-3.259-.014-3.667-.072-4.947-.196-4.354-2.617-6.78-6.979-6.98-1.281-.059-1.69-.073-4.949-.073zm0 5.838c-3.403 0-6.162 2.759-6.162 6.162s2.759 6.163 6.162 6.163 6.162-2.759 6.162-6.163c0-3.403-2.759-6.162-6.162-6.162zm0 10.162c-2.209 0-4-1.79-4-4 0-2.209 1.791-4 4-4s4 1.791 4 4c0 2.21-1.791 4-4 4zm6.406-11.845c-.796 0-1.441.645-1.441 1.44s.645 1.44 1.441 1.44c.795 0 1.439-.645 1.439-1.44s-.644-1.44-1.439-1.44z"></path></svg></a></div></div></div><div><h4 class="text-sm font-semibold text-charcoal-950 uppercase tracking-wider mb-4">Products</h4><ul class="space-y-3"><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/medman">Medman</a></li><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/agentic-harnessing">Agentic Harnesses</a></li><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/agentskills">Agent Skills</a></li><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/agentbuilder">Agent Builder</a></li></ul></div><div><h4 class="text-sm font-semibold text-charcoal-950 uppercase tracking-wider mb-4">Platform</h4><ul class="space-y-3"><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/agentskills/connections/mcp-server-directory">MCP Connectors</a></li><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/agentskills/connections/plugins-directory">Claude Plugins</a></li><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/agentskills/use-cases">Use Cases</a></li><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/agenttesting">Agent Testing</a></li><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/agentinsights">Agent Insights</a></li><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/changelog">Changelog</a></li></ul></div><div><h4 class="text-sm font-semibold text-charcoal-950 uppercase tracking-wider mb-4">Resources</h4><ul class="space-y-3"><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/blog">Blog</a></li><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/learn">Learn / Academy</a></li><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/showcase">Showcase</a></li><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/agentskills/library">Skills Library</a></li><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/claude-skills">Claude Skills</a></li><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/medman/payers">Payer Directory</a></li><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/medman/compare">Compare Options</a></li><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/help">Help Center</a></li></ul></div><div><h4 class="text-sm font-semibold text-charcoal-950 uppercase tracking-wider mb-4">Industries</h4><ul class="space-y-3"><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/medman">Healthcare â Medman</a></li><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/legal">Legal â Debt & M&A</a></li></ul></div><div><h4 class="text-sm font-semibold text-charcoal-950 uppercase tracking-wider mb-4">Company</h4><ul class="space-y-3"><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/about">About</a></li><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/pricing">Pricing</a></li><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/newsroom">Newsroom</a></li><li><a class="text-sm text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/security">Security & Governance</a></li></ul></div></div><div class="pt-8 border-t border-gray-200 flex flex-col md:flex-row justify-between items-center gap-4"><p class="text-sm text-charcoal-950">© <!-- -->2026<!-- --> Agentman (Chain of Agents, Inc.). Crafted with â¤ï¸ in SF</p><div class="flex gap-6 text-sm"><a class="text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/privacy">Privacy</a><a class="text-charcoal-950 hover:text-agentman-600 transition-colors duration-200" href="/terms">Terms</a></div></div></div></footer></main></div></div></div>
10<script id="__NEXT_DATA__" type="application/json">{"props":{"pageProps":{"posts":[{"slug":"prior-auth-maze-thirty-rulebooks-six-front-doors","title":"Newsletter #7: The Prior Auth Maze You Built Yourself: Thirty Rulebooks, Six Front Doors, and No Map","date":"2026-09-02T16:00:00.000Z","readingTime":"15 min read","excerpt":"One continuous glucose monitor, three different submission channels, depending on the payer. After nine years, our practice had six front doors for prior auth and no map â and most of what makes a PA succeed lives in one coordinator's head. Here's the whiteboard workflow we drew before writing a line of code.","category":"Healthcare","segments":["Healthcare","Newsletter"],"tags":["Newsletter","The Agentic Practice","Agentic Healthcare","Prior Authorization","Inbox Triage","Eligibility Verification","Denial Management","Revenue Cycle Management","Independent Practice","Practice Operations"],"author":{"name":"Sachin Gangupantula","avatar":"/images/sachin-gangupantula.jpeg"},"heroImage":"/images/blog/News7_image1.png","heroImageAlt":"The Agentic Practice Issue 7 cover: The Prior Auth Maze You Built Yourself â thirty rulebooks, six front doors, and no map","faq":null},{"slug":"why-practices-wait-prior-authorization","title":"Should You File a Prior Auth Before the Pharmacy Asks? 83% of Prescribers Say No.","date":"2026-08-10T15:00:00.000Z","readingTime":"10 min read","excerpt":"CoverMyMeds publishes that patients get their medications 13.2 days sooner when prior auths are started at the point of prescribing. Their own survey found only 17% of providers do it. The gap isn't ignorance â there's a real argument for waiting that nobody has written down.","category":"Healthcare","segments":["Healthcare"],"tags":["Healthcare","Prior Authorization","Practice Operations","Independent Practice","Revenue Cycle Management"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/prior-auth-wait-vs-file-hero.jpg","heroImageAlt":"Two paths drawn as thin charcoal lines from a single starting point. The upper path runs straight to a tan form outline with several empty fields. The lower path curves through a small terracotta circle before reaching an identical form whose fields are already filled with short charcoal lines â the same destination, reached with different amounts of work.","faq":[{"question":"Should you start a prior authorization before the pharmacy requests one?","answer":"It depends on what you are optimizing for, and reasonable people disagree. CoverMyMeds publishes research that patients access medications 13.2 days sooner when requests are initiated prospectively at the provider's office. But their own provider survey found only 17% of providers do this, and the reason is that a payer-generated request arrives with the correct form already selected and fields pre-populated from live claim data, while a prospectively filed request requires staff to choose the form themselves from benefit identifiers they may not have current."},{"question":"Why do most providers wait for the pharmacy to start a prior authorization?","answer":"Because waiting produces a better-formed request with less work. When a pharmacy claim rejects, the system selects the payer-specific form from the plan and rejection data and auto-populates fields from the claim. Filing prospectively means selecting the form manually using the BIN, PCN, and group numbers from the patient's pharmacy benefit card, which practices frequently do not have in current form, particularly at the start of a plan year when coverage changes."},{"question":"What does the 13.2 days sooner statistic actually measure?","answer":"It measures time to medication access among prescriptions that went through prospective prior authorization compared with retrospective. It does not measure the staff cost of prior authorizations filed for prescriptions that would have adjudicated without one, or for prescriptions a patient never picked up. Those cases produce no time saving because there was no delay to avoid, and the work still had to be done."}
10,{"question":"How is the correct prior authorization form selected?","answer":"CoverMyMeds instructs users to enter the BIN, PCN, and RxGroup numbers from the patient's insurance card, stating those numbers yield the most accurate results and that if more than one form populates, the user should choose the one that best fits the circumstance. When a request is generated from a pharmacy claim rejection instead, the system selects the form using the plan and rejection data rather than requiring a person to choose."},{"question":"What happens if you submit the wrong prior authorization form?","answer":"Staff who do this work report that a mismatched form is rejected without the clinical content being considered, requiring the request to be restarted on the correct form. CoverMyMeds does not publish a specific consequence for form mismatch, and industry discussion of BIN, PCN, and group errors focuses on claim-level misrouting and denials, so treat the operational account as practitioner experience rather than vendor-documented behavior."},{"question":"When does filing a prior authorization prospectively make sense?","answer":"When you already know the drug requires authorization for that plan, you have current benefit identifiers, and the delay genuinely harms the patient. Medications with near-universal authorization requirements are the clearest case, since the probability the request turns out to be unnecessary is low. The calculation changes when a practice cannot confirm current coverage details or when the medication may not require authorization for that particular plan."},{"question":"Can prior authorization work be automated?","answer":"The preparation can be. An agent can open a payer-generated request by its access key, verify which fields arrived pre-populated, fill remaining fields from the chart, and save a draft for review. What should not be automated is submission, because a submitted request is a clinical and financial assertion to a payer. Our agents are prohibited from submitting and stop to report when a required clinical value is missing rather than guessing."}]},{"slug":"refill-and-prior-auth-same-fax","title":"The Refill and the Prior Auth Are the Same Fax, Arriving Twice","date":"2026-08-10T14:00:00.000Z","readingTime":"9 min read","excerpt":"A prescription goes out. The pharmacy tries to fill it. The claim rejects. That rejection generates a prior authorization request that arrives back at the practice as a fax â the same clinical event, now a second document, landing in the same pile as everything else.","category":"Healthcare","segments":["Healthcare"],"tags":["Healthcare","Prior Authorization","Practice Operations","Inbox Triage","Independent Practice","Agentic Healthcare"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/refill-pa-same-fax-hero.jpg","heroImageAlt":"Two tan documents with folded corners sit side by side, joined beneath by a single thin charcoal line that loops from one to the other and back, with a small terracotta circle at the loop's midpoint â two separate pages describing one underlying event.","faq":[{"question":"Who initiates a prior authorization, the doctor or the pharmacy?","answer":"Most often the pharmacy, indirectly. CoverMyMeds documents that when a pharmacy's claim is rejected because prior authorization is required, the pharmacy's dispensing system can send the rejection to CoverMyMeds, which uses the plan and rejection data to select the proper form, auto-populate fields, and share the request electronically with the prescriber. CoverMyMeds' own provider survey found only 17% of providers start a prior authorization at the point of prescribing, meaning the large majority of requests originate downstream of a rejected claim."},{"question":"What is a CoverMyMeds key and how does a practice get one?","answer":"It is a short alphanumeric code that uniquely identifies a specific prior authorization request. When a pharmacy starts a request for one of your patients, the practice receives a fax containing that key, and may also get an email notification. Staff enter the key along with the patient's last name and date of birth to open the request, which arrives partly pre-populated from the pharmacy's claim data."},{"question":"Why does the same prescription generate two separate documents?","answer":"Because the refill authorization and the prior authorization are handled by different systems at different moments. The pharmacy first asks the practice to authorize the refill, and separately, when the pharmacy attempts to adjudicate the claim, a rejection can generate a prior authorization request that returns to the practice as its own fax. One
10clinical decision produces two pieces of paper, often days apart, arriving through the same channel as everything else."},{"question":"How do practices tell a refill request from a prior authorization request?","answer":"Traditionally by opening the document and reading it, because both arrive as faxes in the same queue with no distinguishing envelope. The difference matters because they carry different urgency and different work. Automated classification can separate them by matching layout and content patterns, such as the presence of an access key, the sending pharmacy's form structure, or headers indicating a response is required."},{"question":"Why do prior authorization requests get missed in a medical practice?","answer":"Because they arrive undifferentiated among dozens of other faxes, and nothing about their arrival signals urgency. A prior authorization sitting unopened means a patient is not receiving a medication their physician already prescribed, and if the request is a denial with an appeal window, the clock is running while the document sits. The failure is rarely a decision to deprioritize it and almost always a failure to notice it."},{"question":"Does the prior authorization for a CGM or DME work the same way as for a drug?","answer":"No. Under a pharmacy benefit, requests typically route through electronic prior authorization tools such as CoverMyMeds. Under a medical or durable medical equipment benefit, Medicare documentation identifies the supplier rather than the prescriber as the requester, and some payers state explicitly that the prescribing provider does not submit the request at all. Which path applies depends on how the specific plan assigns the benefit, which is why the same device can follow different routes for different patients."},{"question":"Can an AI agent handle prior authorization paperwork?","answer":"It can prepare the work, and that is where we draw the line. An agent can open a request by key, check which fields the payer already populated, fill in what is missing from the chart, and save a draft. Our agents are explicitly prohibited from submitting to the plan or routing to the prescriber, and are built to interpret even a direct instruction to submit as an instruction to save the draft, so a person reviews and sends every request."}]},{"slug":"third-of-refill-requests-are-not-refills","title":"A Third of Refill Requests Aren't Refill Requests","date":"2026-08-09T15:00:00.000Z","readingTime":"9 min read","excerpt":"In a year-long study at a federally qualified health center, 36% of refill requests carried a clinical finding â a patient overdue for labs, a dose that should have changed, a drug that should have stopped. The more troubling number: of the drug therapy problems found, 55% were never addressed.","category":"Healthcare","segments":["Healthcare"],"tags":["Healthcare","Practice Operations","Clinical Quality","Independent Practice","Agentic Healthcare"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/third-of-refills-clinical-hero.jpg","heroImageAlt":"Ten small outlined pill-shaped tiles arranged in a grid on cream; six are plain tan, and four are terracotta, each marked with a small charcoal symbol â a calendar, a test tube, an upward arrow, and a cross â showing that a substantial share of routine-looking requests carry a finding.","faq":[{"question":"What percentage of prescription refill requests require clinical intervention?","answer":"In a year-long study at a federally qualified health center published in the Journal of Primary Care and Community Health, 36.1% of pharmacist-managed refill encounters resulted in at least one clinical intervention. The study cites corroborating research by Nguyen and Zare that found interventions needed in 42% of pharmacist-managed refills across more than 5,000 requests over two years, so the finding appears consistent across settings."},{"question":"What kinds of problems are found during prescription refill review?","answer":"The most common findings are administrative-clinical rather than dramatic. In the published study, 32.9% of encounters involved a patient due for a follow-up visit and 31.9% a patient due for laboratory monitoring. Beyond those, 12.6% surfaced a drug therapy problem, 11.9% found a discrepancy between the pharmacy request and the chart medication list, and 7.2% identified a patient who should be referred to another clinical service."},{"question":"What is a drug therapy problem in a refill request?","answer":"It is a medication issue that makes the request as written clinically inappropriate. In the published study the most common categories were dosage too low at 24.1%, requires a different drug product at 21.7%, dosage too high at 20.5%, and unnecessary drug therapy at 20.5%. Concrete examples given include aspirin continued for cardiovascular prophylaxis in low-risk patients, therapeutic duplications, and diabetes medications that current guidelines suggest replacing."},{"question":"Do clinicians act on medication problems found during refill review?","answer":"Often not. The study followed up at three months on the 83 drug therapy problems identified and found that 37.3% had been addressed or resolved, 55.4% had not been addressed at all, and the remainder were indeterminate. The clinical work of finding the problem was completed and documented, and more than half of it produced no action, which makes tracking to closure at least as important as detection."},{"question":"Should prescription refills be automatically approved?","answer":"Blanket auto-approval optimizes for the wrong outcome. The refill queue functions as one of the few moments a chronic medication regimen gets reviewed outside a scheduled visit, and the published evidence indicates a meaningful share of requests carry a finding. A system that approves the routine-looking ones faster without examining them removes the occasion for looking, so the appropriate goal is surfacing the requests that need a human rather than minimizing the number that reach one."},{"question":"Which medications should be excluded from refill protocols?","answer":"The published protocol excluded controlled substances, medications for acute c
10onditions, pain medications including NSAIDs and muscle relaxants, proton pump inhibitors, and antibacterial and antifungal medications. It also required that patients be established adults with an in-office or telemedicine encounter within the preceding twelve months. Requests outside those criteria were forwarded to the primary care provider untouched rather than handled by protocol."},{"question":"How much clinician time do prescription refills actually take?","answer":"Less than commonly claimed, and the widely circulated figures are poorly sourced. The AMA has published a two-hours-per-day savings claim, but it is attributed to an individual's assertion and personal experience rather than a study, and it describes synchronized annual prescribing rather than automation. In the peer-reviewed study, pharmacists handled 1,683 refill requests in roughly 3.2 hours per week, averaging about 10 requests per hour."}]},{"slug":"refill-request-is-not-a-task","title":"A Refill Request Isn't a Task. It's an Unclassified Trigger.","date":"2026-08-09T14:00:00.000Z","readingTime":"12 min read","excerpt":"Every vendor selling refill automation answers the same question: what percentage can you auto-approve? It's the wrong question. A refill request has no identity until someone reads it â and in our production triage system, it can turn out to be any of six different jobs.","category":"Healthcare","segments":["Healthcare"],"tags":["Healthcare","Practice Operations","Inbox Triage","Prior Authorization","Independent Practice","Agentic Healthcare"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/refill-unclassified-trigger-hero.jpg","heroImageAlt":"A single tan document at the left joined by thin charcoal lines that fan out to six small outlined shapes: a checkmark, a calendar, a vial, a clipboard form, two mismatched rectangles, and an arrow turning away â one arriving request resolving into six different kinds of work.","faq":[{"question":"What percentage of prescription refills can be automated?","answer":"This is the wrong question to optimize against, and the figures circulating online are unreliable. Many blogs attribute precise automation rates to MGMA that do not appear in any MGMA publication we could locate. More importantly, an auto-approval rate only describes requests where approval was the correct answer. It says nothing about requests that should have triggered a lab order, a dose change, a prior authorization, or an appointment, which is where both the clinical risk and the staff time actually sit."},{"question":"What are the different outcomes of a prescription refill request?","answer":"At least six. Straightforward approval; approval only after the patient comes in for a visit or labs; a clinical decision that the medication should be changed, adjusted, or stopped; a prior authorization that must be filed before the pharmacy can dispense; a reconciliation problem where the pharmacy's request does not match the chart; and a redirect because the request belongs to a different prescriber or is a duplicate. Each is a different workflow with a different owner."},{"question":"Why does a refill request turn into a prior authorization?","answer":"Because the prior authorization is usually generated by the pharmacy's failed fill attempt, not by the practice. The prescription is sent, the pharmacy runs it against the patient's benefit, the claim rejects, and that rejection generates an electronic prior authorization request that arrives back at the practice as a fax. The refill and the prior authorization are the same clinical event arriving as two separate documents, often days apart."},{"question":"How does an AI inbox triage agent classify a refill request?","answer":"By reading the document and matching it against known patterns: the sending pharmacy's form layout, headers such as RESPONSE REQUESTED, the presence of a CoverMyMeds key, or whether a supply order carries a CMN form number. Our production classifier assigns a category, a sub-type, a priority, and a confidence score, and extracts patient identifiers and medication details into structured fields. When confidence falls below a threshold it declines to classify rather than guessing."},{"question":"Can an AI agent approve prescription refills automatically?","answer":"It should not, and ours does not. Classification determines what kind of request arrived and how urgent it is, then routes it to a person. Downstream agents can prepare work such as filling out a prior authorization form, but the terminal action is saving a draft for human review rather than submitting it. The clinical decision and the submission both stay with a licensed human."},{"question":"What should never be automated in a refill workflow?","answer":"Controlled substances are the clearest boundary, because they carry prescriber-identity, PDMP-check, and refill-limit requirements that sit outside routine protocol handling. Published pharmacist-managed refill protocols also exclude medications for acute c
10onditions, pain medications, and antibiotics, because a refill request for an acute course usually signals that something did not resolve rather than that a renewal is due."},{"question":"Why do refill requests arrive through so many channels?","answer":"Because every party involved uses a different one. The pharmacy faxes a refill authorization form, the patient leaves a voicemail or sends a portal message, the benefit manager emails a prior authorization notification, and durable medical equipment suppliers use their own ordering platforms. The same underlying request for the same medication can arrive three separate times through three separate channels, which is why deduplication has to happen before anyone counts volume."}]},{"slug":"denial-report-cannot-see-denials","title":"Your Denial Report Can't See Denials. Here's the Proof.","date":"2026-08-08T14:00:00.000Z","readingTime":"12 min read","excerpt":"We analyzed one practice's denials and found $98 worth worth appealing. Then we read the actual remittances and found $5,413 â in the same month, at the same practice. The first method was structurally blind, and it's the method most denial reports use.","category":"Healthcare","segments":["Healthcare"],"tags":["Healthcare","Revenue Cycle Management","Denial Management","Practice Operations","Independent Practice"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/denial-report-blind-hero.jpg","heroImageAlt":"Six bars on a baseline inside a thin rectangular report frame. Five are short and tan; one, in terracotta, is several times taller and breaks straight through the top edge of the frame â the denials a payment-derived report cannot contain or show.","faq":[{"question":"Why do denial reports miss denials?","answer":"Many denial reports are built from posted payments â they read the adjustment reason code attached to each payment record. That works for a claim the payer paid something on and reduced. It fails completely for a claim the payer denied outright, because a fully denied line pays $0.00, posts no payment record, and therefore carries no reason code for the report to read. The denials that cost you the most are the ones the report is structurally unable to see."},{"question":"What is the difference between CARC and RARC codes?","answer":"CARC is the claim adjustment reason code, which states why an amount was not paid. RARC is the remittance advice remark code, which usually says what the payer wants you to do about it. CO-251 alone means missing information. CO-251 paired with RARC N479 means the primary payer's explanation of benefits was not attached, which is an instruction rather than a dispute. The remark code is where the fix usually lives, and it only appears in the remittance."},{"question":"Where do I find my practice's real denial data?","answer":"In the electronic remittance advice, the 835 file the payer sends with each check. That is the only record that carries the reason code for a fully denied line, because it describes the adjudication rather than the payment. Some practice management systems expose remittances through an API, some only through a downloadable report, and retention windows vary. Anything older than your system's retention window generally has to come from your clearinghouse."},{"question":"How much are denials actually costing a small practice?","answer":"It varies enormously by payer mix and specialty, and any vendor quoting one number across all practices is quoting a best case. In the single month we examined at one small specialty practice, 139 fully denied lines represented $7,659 in billed charges, of which we classified $5,413 as carrying a fixable reason. What matters more than the figure is that it was concentrated: one payer accounted for more than half, and the largest patterns repeated on multiple service dates within the month."},{"question":"Is it better to appeal denials or prevent them?","answer":"Prevent them, where the denial is caused by a rule you can check before submission. A prevented denial costs nothing to fix and never enters anyone's work queue, while an appealed denial costs staff time even when it succeeds. In the month we analyzed, three of the four largest patterns were pre-submission rule failures rather than genuine adjudication disputes, and they recurred on a schedule. Appeals still matter for what prevention cannot catch."},{"question":"How many denied claims are never resubmitted?","answer":"A large share. Health Affairs, analyzing 270 million Medicare Advantage claim submissions, found that 60% of denied claims were resubmitted â meaning 40% never were. Of those resubmitted, tw
10o-thirds were overturned. Industry sources put the never-resubmitted figure across payers even higher. The money is not usually lost to failed appeals; it is lost to appeals nobody filed."},{"question":"What should a denial report actually show?","answer":"Every denied line including the ones that paid zero, the CARC and RARC together, the payer, the service date, and the dollar amount, grouped so repeating patterns are visible. It should also separate genuinely non-appealable adjustments such as contractual write-offs from refusals with a fixable reason, and it should say which bucket unknown codes went into. A report that collapses everything into one contractual line is not wrong so much as incomplete in the specific place that costs money."}]},{"slug":"security-questions-before-signing-ai-vendor","title":"21 Questions to Ask an AI Vendor Before You Sign","date":"2026-08-08T00:00:00.000Z","readingTime":"9 min read","excerpt":"A security questionnaire a small practice can actually run â no procurement team required. Copy the list, send it, and read the answers for artifacts rather than adjectives. We answer all 21 ourselves at the bottom.","category":"Healthcare","segments":["Healthcare"],"tags":["Healthcare","Compliance","Vendor Evaluation","HIPAA","Practice Operations"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/security-questions-ai-vendor-hero.jpg","heroImageAlt":"A tall tan checklist document outlined in charcoal with a column of short horizontal lines beside small square checkboxes, three of them marked with terracotta checks, and a small terracotta magnifying glass resting over the lower portion of the list.","faq":[{"question":"What security questions should a small practice ask an AI vendor?","answer":"Start with six that filter fast: will you sign a BAA, who are your subprocessors, is our data used to train models, what is your breach notification timeline in hours, can you show us a single agent action from six months ago, and what happens to our data when we leave. A vendor who answers those six with documents rather than adjectives is worth the longer conversation, and one who cannot answer them will not improve under further questioning."},{"question":"Do I need a procurement team to evaluate an AI vendor?","answer":"No. Most of what a formal security review produces can be obtained by sending a written list of questions and reading the answers carefully. The advantage a procurement team has is time and pattern recognition, not access â every artifact that matters is something a vendor will hand to any serious buyer who asks. An afternoon spent on the BAA and the subprocessor list covers most of the real risk."},{"question":"What is a red flag in a vendor security answer?","answer":"Adjectives where an artifact belongs. Bank-grade security, enterprise-ready, fully compliant, and military-grade encryption are all phrases that describe nothing checkable. The specific pattern to watch for is a claimed certification that cannot be produced on request, because a real report is a file a vendor sends under NDA as a routine part of procurement."},{"question":"Should I ask for a SOC 2 report or is the badge enough?","answer":"Ask for the report. A badge does not tell you the type, the observation period, the scope, or whether there were exceptions, and all four determine whether the attestation covers anything you care about. Vendors with genuine Type II reports share them under NDA without friction, so a request that meets resistance has told you something useful."},{"question":"What should happen to our data when we stop using an AI vendor?","answer":"It should be returned or destroyed, and the contract should say which. HHS sample BAA provisions call for return or destruction of PHI at termination where feasible, with protections continuing for anything retained. Ask specifically about backups, prompt logs, and derived data, since those copies are the ones most often missed when a deletion request is processed."},{"question":"How do I verify a vendor's claims rather than trusting them?","answer":"Ask for the artifact behind each claim and check one of them yourself. A SOC 2 claim becomes a report with a date range, an audit-trail claim becomes a walkthrough of one real action from months ago, a no-training claim becomes a clause number, and a subprocessor claim becomes a list. Any claim that cannot be converted into a document or a demonstration should be treated as marketing."},{"question":"Is it reasonable to ask a small AI vendor for all of this?","answer":"Yes, though the answers will differ from what a large vendor gives you and that is not automatically disqualifying. A young company may lack a SOC 2 report while having a real risk analysis, a signed BAA, and clear data handling. What matters is whether the answers are specific and whether the gaps are stated plainly rather than papered over with badges."}]},{"slug":"can-ai-read-phi-data-handling-rules","title":"Can AI Read PHI? The Data-Handling Rules That Actually Apply","date":"2026-08-07T23:30:00.000Z","readingTime":"8 min read","excerpt":"Yes â with a BAA and the right safeguards. The harder questions are where the data travels during inference, how little of it the agent actually needs, and what ends up sitting in prompt logs nobody thought to check.","category":"Healthcare","segments":["Healthcare"],"tags":["Healthcare","HIPAA","AI Governance","Compliance","Data Protection"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/can-ai-read-phi-hero.jpg","heroImageAlt":"A tall tan document with several lines of text, most of them covered by solid charcoal redaction bars, and a single terracotta-outlined magnifying glass hovering over the two lines left visible â an agent seeing only the minimum necessary fields.","faq":[{"question":"Can an AI agent legally read protected health information?","answer":"Yes, provided the vendor is under a business associate agreement and the arrangement meets HIPAA's Privacy and Security Rule requirements. Nothing in HIPAA prohibits software from processing PHI, and practices have used software to handle patient data for decades. What HIPAA requires is that the entity handling it has appropriate safeguards, a contract defining permitted uses, and accountability if something goes wrong."},{"question":"Does the minimum necessary standard apply to AI agents?","answer":"Yes, in the same way it applies to staff. The Privacy Rule limits uses and disclosures of PHI to the minimum necessary to accomplish the purpose, and an agent verifying insurance eligibility does not need the full clinical note to do it. In practice this means asking a vendor which fields the agent actually receives, because the easy engineering choice is to send everything and filter later."},{"question":"Where does PHI go when an AI agent processes it?","answer":"That depends on the architecture, and it is the question buyers most often skip. Data may pass from the vendor's application to a model provider for inference, to a cloud host for storage, and potentially to telephony or document-processing services. Each hop is a place PHI exists, each needs its own agreement, and the only way to know the real path is to ask for a named subprocessor list rather than a reassurance."},{"question":"Is PHI stored in AI prompt logs?","answer":"Often, unless someone deliberately prevented it. Debugging infrastru
10cture, prompt traces, and vector stores routinely retain the content sent through them, and these systems are frequently overlooked when retention and deletion policies are written. Ask specifically whether prompts containing PHI are logged, how long those logs live, who can read them, and whether they are included when you request deletion."},{"question":"What is the difference between de-identified data and PHI?","answer":"De-identified data is health information from which identifiers have been removed under one of two HIPAA methods: expert determination, where a qualified statistician concludes re-identification risk is very small, or safe harbor, which requires removing eighteen specified identifier types. Properly de-identified data is no longer PHI and falls outside HIPAA. Partial masking that leaves dates, ZIP codes, or record numbers in place does not meet either standard."},{"question":"Should staff be able to see everything an AI agent sees?","answer":"Not necessarily, and role-based access should apply to agent outputs the way it applies to any other system. The Security Rule requires access controls allowing only authorized persons to reach ePHI, and that principle extends to operational dashboards showing agent activity. A support view that exposes complete records to anyone troubleshooting is a wider surface than most practices intend."},{"question":"What encryption does HIPAA require for AI systems?","answer":"HIPAA does not mandate a specific algorithm. The Security Rule sets standards while categorizing some implementation specifications as required and others as addressable, and encryption is addressable â meaning an entity must implement it where reasonable and appropriate, or document why an equivalent alternative was chosen. In practice, encryption in transit and at rest is the expected baseline, and a vendor without it should be able to explain why."}]},{"slug":"soc-2-for-ai-vendors","title":"SOC 2 for AI Vendors: What It Covers and What It Doesn't","date":"2026-08-07T23:00:00.000Z","readingTime":"9 min read","excerpt":"A SOC 2 report is real evidence â but the vendor helped choose what was examined, and the badge on the website tells you none of that. What to ask for, how to read the report, and why 'SOC 2 ready' means nothing at all.","category":"Healthcare","segments":["Healthcare"],"tags":["Healthcare","Compliance","Vendor Evaluation","AI Governance","Security"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/soc-2-for-ai-vendors-hero.jpg","heroImageAlt":"A tan badge shape on the left with a small terracotta question mark inside it, and on the right a multi-page report document with a visible date range bar and several short lines, connected by a thin charcoal arrow pointing from badge to report.","faq":[{"question":"What is the difference between SOC 2 Type I and Type II?","answer":"Type I says controls were suitably designed at a single point in time. Type II says an auditor observed those controls operating across a period, typically three to twelve months. Type II is substantially stronger because a control that exists on paper and a control that functions every day are different things. When a vendor says SOC 2 without specifying, ask which one and ask for the observation period."},{"question":"Does SOC 2 mean a vendor is secure?","answer":"It means an independent auditor examined the controls the vendor put in scope against selected trust services criteria and reported what they found. That is meaningful evidence, but scope is chosen with the vendor's participation, so a narrow scope can produce a clean report that covers little of what you care about. Read the scope section and the exceptions before treating the report as an answer."},{"question":"What does 'SOC 2 ready' mean?","answer":"It is not an attestation. It generally means a company believes it has implemented the controls a SOC 2 audit would examine, but no auditor has issued a report. The phrase has no standard definition and no third party stands behind it, so treat it as a statement of intent. The verifiable question is whether a report exists, from which firm, covering what period."},{"question":"Is SOC 2 required for HIPAA compliance?","answer":"No. SOC 2 is a voluntary attestation developed by the AICPA, while HIPAA is federal law enforced by the HHS Office for Civil Rights. Neither substitutes for the other. A vendor handling PHI needs a business associate agreement regardless of whether it holds a SOC 2 report, and a SOC 2 report is not a finding about HIPAA compliance."},{"question":"What should I ask a vendor about their SOC 2?","answer":"Ask four things: which type, which observation period, which trust services criteria were in scope, and whether there were exceptions. The
10n ask for the report itself under NDA rather than accepting a summary or a logo. A vendor with a genuine Type II report will hand it over as a routine part of procurement; hesitation at this step is itself information."},{"question":"Are exceptions in a SOC 2 report a dealbreaker?","answer":"Usually not, and a report with zero exceptions across a long observation period is sometimes a sign of narrow scope rather than flawless operation. What matters is what the exception was, whether it touched data like yours, and what the vendor did about it. A vendor who explains an exception clearly is often a safer bet than one whose report has nothing in it to discuss."},{"question":"Can a small vendor be trusted without SOC 2?","answer":"Yes, depending on what they can show you instead. Audits are expensive and early-stage companies frequently have real controls before they have a report. The substitutes worth accepting are concrete: a documented risk analysis, a signed BAA where PHI is involved, a penetration test summary, a named subprocessor list, and a clear breach notification commitment. What should not be accepted is a badge implying an attestation that does not exist."}]},{"slug":"hipaa-compliant-ai-agents-baa-phi-audit-trails","title":"HIPAA-Compliant AI Agents: BAAs, PHI Handling, and Audit Trails","date":"2026-08-07T22:30:00.000Z","readingTime":"9 min read","excerpt":"An AI agent that touches patient data is a business associate, with direct legal liability. Here's what the BAA has to cover, which AI-specific questions older templates miss, and what an audit trail needs to contain when an agent acted instead of a person.","category":"Healthcare","segments":["Healthcare"],"tags":["Healthcare","HIPAA","AI Governance","Compliance","Practice Operations"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/hipaa-ai-agents-baa-audit-hero.jpg","heroImageAlt":"Three elements in a row joined by thin lines: a tan document with signature marks, a horizontal chain of three linked tiles, and a block of stacked terracotta dashes beside a vertical rule â the signed agreement, the subprocessor chain it binds, and the audit log that records what happened.","faq":[{"question":"Is an AI vendor a business associate under HIPAA?","answer":"Yes, if it creates, receives, maintains, or transmits protected health information on behalf of a covered entity. HHS defines a business associate as a person or entity outside your workforce performing functions involving access to PHI, and that definition also covers subcontractors handling PHI on behalf of another business associate. An AI agent reading charts, verifying eligibility, or handling patient messages fits squarely inside it."},{"question":"What does a BAA need to cover that is specific to AI?","answer":"Three things older templates often miss. First, whether PHI may be used to train or fine-tune models, which should generally be prohibited outright. Second, the subprocessor chain, because AI vendors typically route data through model providers and cloud hosts that each need their own agreements. Third, retention inside model and logging infrastructure, since prompt logs and vector stores can hold PHI in places a traditional retention clause never contemplated."},{"question":"Should PHI ever be used to train an AI model?","answer":"Not without explicit contractual permission, and for most practices the answer should simply be no. A business associate may use or disclose PHI only as permitted by its agreement or as required by law, so silence in the contract is not permission â but silence is also a gap you do not want to discover after the fact. Ask for an explicit written prohibition rather than a verbal assurance."},{"question":"What should an AI agent's audit trail contain?","answer":"Enough to reconstruct why the agent did what it did, not just that it did it. That means a timestamp, the identity of the agent and its version, the input it acted on, the specific action taken, the output produced, and the human who reviewed or approved it where applicable. When software rather than a person takes an action on a record, the version of the logic matters as much as the actor, because reconstructing a decision from six months ago requires knowing which version was running."},{"question":"How long do HIPAA records need to be retained?","answer":"HIPAA requires documentation of policies, procedures, and required actions or assessments to be retained for six years from the later of the date it was created or the date it was last in effect. That applies to the compliance documentation itself. Retention of PHI is governed separately by state law and your own record-retention policy, which is why the BAA needs to state what happens to PHI at termination."},{"question":"Who is liable if an AI agent causes a HIPAA breach?","answer":"The business associate is directly liable under HIPAA and can face civil and, in some cases, criminal penalties. The covered entity remains responsible for obtaining satisfactory assurances that its business associates safeguard PHI, so liability is shared rather than transferred. Contracting with a vendor does not move your obligation to them, which is why the diligence has to happen before deployment."},{"question":"What happens to our PHI when we stop using an AI vendor?","answer":"Whatever the BAA says, which is why you should read that clause before signing. HHS's sample provisions require the business associate to return or destroy all PHI at termination where feasible, and to extend protections to anything retained for as long as it is held. Ask specifically about backups, logs, and any derived data, because those are the copies most often overlooked in a deletion request."}]}
10,{"slug":"is-ai-hipaa-compliant","title":"Is AI HIPAA Compliant? There Is No Such Thing as HIPAA Certified","date":"2026-08-07T22:00:00.000Z","readingTime":"9 min read","excerpt":"HHS does not certify anyone as HIPAA compliant â it says so in writing. Here's what the label actually means when an AI vendor uses it, what to ask for instead, and the one document that carries legal weight.","category":"Healthcare","segments":["Healthcare"],"tags":["Healthcare","HIPAA","AI Governance","Vendor Evaluation","Compliance"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/is-ai-hipaa-compliant-hero.jpg","heroImageAlt":"A tan certificate shape with a rosette seal, outlined in charcoal, with a terracotta line struck through it. Beside it, a plain rounded document with signature lines at the bottom and a terracotta outline â the certificate that does not exist next to the signed agreement that does.","faq":[{"question":"Is there such a thing as HIPAA certification?","answer":"No. There is no government-issued HIPAA certification, and no federal seal or registry exists. HHS states directly that it does not endorse or recognize private organizations' certifications regarding the Security Rule, and that such certifications do not absolve anyone of their legal obligations. Any vendor calling itself HIPAA certified is describing a private company's assessment or a training completion certificate, not a government determination."},{"question":"What does it mean when an AI vendor says it is HIPAA compliant?","answer":"It means the vendor is making a self-assessment. The label carries no government endorsement because no agency pre-clears or certifies products. It may well be accurate â plenty of vendors take their obligations seriously â but it is a claim, not a credential, and it should be treated the same way you would treat any other unverified vendor statement."},{"question":"What should I ask an AI vendor instead of asking if they are HIPAA compliant?","answer":"Ask whether they will sign a business associate agreement, and read it. Then ask for the artifacts behind it: their most recent security risk analysis, a list of subprocessors that will touch PHI, their breach notification timeline, and whether PHI is used to train models. A signed BAA plus documented controls tells you far more than any badge, because the BAA creates enforceable legal obligations and the badge creates none."},{"question":"Does a business associate agreement make a vendor HIPAA compliant?","answer":"It does not make them compliant, but it makes them legally accountable, which is the more useful thing. Under the HITECH Act, business associates are directly liable for Security Rule violations and face civil and, in some cases, criminal penalties. A BAA documents required assurances, commits the vendor to the Security Rule, extends those obligations to their subcontractors, and obligates them to report breaches to you."},{"question":"Is SOC 2 the same as HIPAA compliance?","answer":"No. SOC 2 is a voluntary attestation about a service organization's controls, issued by an accounting firm against criteria the organization itself helps scope. HIPAA is federal law enforced by the HHS Office for Civil Rights. A SOC 2 report is meaningful evidence that a vendor has controls and had them examined, but it is not a HIPAA determination and it does not replace a BAA."},{"question":"Can an AI vendor use PHI to train its models?","answer":"Only if your business associate agreement permits it, and most should not. A business associate may use or disclose protected health information only as permitted by its contract or as required by law. If model training is not addressed explicitly in the BAA, ask for it in writing before any PHI moves. This is one of the few AI-specific questions that older BAA templates often fail to cover."},{"question":"Who is liable if an AI vendor mishandles PHI?","answer":"Both parties, in different ways. The business associate is directly liable under HIPAA for its own violations. But the covered entity remains responsible for obtaining satisfactory assurances that its business associates safeguard PHI appropriately, so using a vendor that turns out to be careless does not transfer your obligation away. That is
10why the diligence happens before deployment rather than after a breach."}]},{"slug":"integrated-agents-vs-point-solutions","title":"One Vendor or Five? What Buying Your Second Agent Teaches You","date":"2026-08-07T20:00:00.000Z","readingTime":"9 min read","excerpt":"Start with one workflow â that advice is right, and it still works. The trap is choosing your first agent without asking what the second one will cost you. Here's what breaks when eligibility, voice, and prior auth come from three different vendors.","category":"Healthcare","segments":["Healthcare"],"tags":["Healthcare","Revenue Cycle Management","Vendor Evaluation","Practice Operations","Independent Practice"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/integrated-agents-vs-point-solutions-hero.jpg","heroImageAlt":"Split in two by a thin vertical line. On the left, five tan tiles scattered with no lines joining them, above a knot of tangled threads. On the right, one terracotta-outlined panel holding the same five tiles in a neat row, each joined to the next by a short straight line â five disconnected vendors beside one integrated system.","faq":[{"question":"Should I buy all my AI agents from one vendor?","answer":"Not necessarily at the start, but you should choose your first vendor knowing what your second purchase will look like. Starting with one workflow is still the right approach â it proves value before you commit further. The mistake is picking a first agent from a vendor that only sells that one thing, because the expansion decision then costs you a second procurement, a second integration, and a permanent handoff between two systems that were never designed to talk."},{"question":"What actually breaks when agents come from different vendors?","answer":"The handoffs. An eligibility check that discovers a plan requires prior authorization has to reach the prior auth workflow; if those are separate products, a human moves that information. A denial that traces back to a missed eligibility detail cannot automatically inform the eligibility agent. A patient call about a referral has no view of the fax that arrived that morning. Each seam is a place where context is lost and a staff member becomes the integration layer."},{"question":"Isn't it better to buy the best product for each workflow?","answer":"It is a real tradeoff, and for some practices best-of-breed wins. If one workflow dominates your pain and a specialist vendor is clearly stronger at it, buy that. The calculation changes when you automate a second and third workflow, because the value of connected context grows with each one while the coordination cost grows too. Best-of-breed makes most sense when a workflow is genuinely isolated from the others."},{"question":"How much does it cost to add a second vendor?","answer":"More than the license. Each additional vendor means a separate evaluation, a separate security and HIPAA review, a separate EHR integration, a separate support relationship, separate invoices, and separate staff training. Then there is the ongoing cost that has no line item: someone in your practice becomes responsible for moving information between systems that do not share it."},{"question":"What should I ask a vendor before buying my first agent?","answer":"Ask what else they run, and how their agents share context. Specifically: if the eligibility agent finds that a plan requires prior authorization, does the prior auth workflow see it automatically? If a denial arrives, does anything feed back into how future checks are run? Can the voice agent see what arrived by fax this morning? A vendor with one product will answer these in terms of integrations and APIs, which means your staff or your IT budget closes the gap."},{"question":"Does starting with one workflow still make sense?","answer":"Yes, and we would give the same advice today. Deploy one agent, measure it against a baseline you recorded beforehand, and expand only after it works. The practices that try to automate everything at once run into change management resistance. Starting narrow is about sequencing your deployment, not about limiting who you buy from â those are two different decisions that are easy to confuse."},{"question":"What is an integrated agentic back office?","answer":"A set of agents covering different back-office workflows that share context and hand work to each other without a person in between. Eligibility findings inform prior authorization. Denial patterns inform future eligibility checks. Inbound faxes, voicemails, and portal messages route into the same queues the other agents work from. The distinguishing feature is not breadth of features but whether the agents can see each other's work."}]},{"slug":"assort-health-alternatives","title":"Assort Health Alternatives: What to Consider in 2026","date":"2026-08-07T18:00:00.000Z","readingTime":"13 min read","excerpt":"Assort Health sells annual contracts to multi-provider groups, and for that buyer it's a capable product. If you have fewer than five providers, the more useful question isn't which vend
10or is better â it's whether you're the customer at all.","category":"Healthcare","segments":["Healthcare"],"tags":["Healthcare","Voice AI","Vendor Comparison","Independent Practice","Pricing"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/assort-health-alternatives-hero.jpg","heroImageAlt":"A tan doorway outlined in charcoal with a wide arch, sized so that five small figures fit through it together while a single figure and a pair stand outside at a smaller scale â a vendor minimum drawn as a doorway that only groups of five or more can pass through.","faq":[{"question":"What are the alternatives to Assort Health?","answer":"For healthcare voice agents the closest alternatives are Prosper AI, Infinitus, SuperDial, and Agentman. Prosper targets practices and outpatient groups with specialty-specific workflows. Infinitus and SuperDial sell to enterprise buyers â payers, pharma, RCM companies, and health systems. Agentman differs in scope, running voice alongside eligibility, prior authorization, and inbox triage. As of 7 August 2026, Agentman is the only one of the five publishing a rate on its own site."},{"question":"How much does Assort Health cost?","answer":"Assort does not publish pricing; their site directs you to book a demo. We can only report our own experience. In July 2026 a sales inquiry we sent came back stating that a $75,000 annual minimum commitment was required before an initial call could be scheduled, and that the minimum reflected a floor of five providers. That works out to about $15,000 per provider per year. That is the quote we received in our situation, not Assort's pricing policy â quotes vary by practice size, specialty, and site count."},{"question":"Does Assort Health have a minimum practice size?","answer":"In the quote we received in July 2026, yes â a five-provider floor. We cannot speak to whether that applies to every prospect, and Assort publishes no minimum on its site. But it is worth asking directly if you have fewer than five providers, because a minimum expressed in providers rather than dollars determines whether you are in scope at all, and that is a faster thing to learn in the first email than in the third meeting."},{"question":"Is Assort Health a good choice?","answer":"For a multi-provider group or health system that wants patient-facing voice across scheduling, intake, and referrals, it is a serious option. They state deep EHR integration across 15-plus platforms, coverage of 23-plus specialties, and a platform trained on a large volume of interactions. The friction is that you cannot see pricing without entering a sales process, and the contracts are annual."},{"question":"What is the difference between an annual contract and pay-as-you-go?","answer":"An annual contract fixes your commitment before you have production evidence, which is why vendors like it and why minimums exist. Pay-as-you-go bills for what you use, so a practice can run a small volume, measure whether it worked, and stop if it did not. The tradeoff is real in both directions: annual contracts often carry lower effective rates at volume, while usage billing costs more per unit but removes the risk of paying for a year of something that did not fit."},{"question":"Should a small practice buy a healthcare voice agent at all?","answer":"Only after counting where staff time actually goes. Spend a week logging it. If the phone genuinely dominates, a voice agent is worth pricing. In many practices, though, phone time is one of several sinks alongside eligibility verification, inbound faxes and portal messages, and prior authorization. Voice solves the most audible problem, which is not always the most expensive one."},{"question":"How do I compare healthcare voice AI vendors on price when most publish nothing?","answer":"Ask every vendor the same four questions and get the answers in writing: is hold time billed or only connected time, is there a minimum and what unit is it counted in, what does a failed or dead-end call cost, and is there a setup fee. The unit the minimum is counted in matters most. A minimum in dollars reads as a price; a minimum in providers or sites tells you whether a practice your size is a customer."}]},{"slug":"prosper-ai-alternatives","title":"Prosper AI Alternatives: 5 Healthcare Voice Agent Options (2026)","date":"2026-08-07T16:00:00.000Z","readingTime":"12 min read","excerpt":"Prosper AI is a strong healthcare voice agent, and for some practices it's the right call. Here's an honest look at the alternatives â what each one publishes, who each fits, and where we'd tell you to pick someone else.","category":"Healthcare","segments":["Healthcare"],"tags":["Healthcare","Voice AI","Vendor Comparison","Revenue Cycle Management","Independent Practice"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/prosper-ai-alternatives-hero.jpg","heroImageAlt":"Five telephone handsets in a row on a thin baseline. Four are plain tan and stand alone; the fifth, outlined in terracotta, fans four thin lines out to a document, an envelope, a clipboard with a checkmark, and a stack of cards â a voice agent that also handles the back-office queues the others cannot.","faq":[{"question":"What are the main alternatives to Prosper AI?","answer":"The closest alternatives for healthcare voice automation are Infinitus, SuperDial, Assort Health, and Agentman. Infinitus and SuperDial are aimed at enterprise buyers â payers, p
10harma, RCM companies, and health systems. Assort Health and Prosper both target practices and specialty groups. Agentman differs in scope: voice is one agent among several back-office agents rather than the whole product. As of 7 August 2026, Agentman is the only one of the five that publishes a rate on its own site."},{"question":"How much does Prosper AI cost?","answer":"Prosper does not publish pricing. There is no pricing page on getprosper.ai, and their own blog describes their model as usage-based with a custom quote, noting that pricing is not public so early budgeting requires a sales conversation. Third-party sites circulate per-provider estimates, but those are not Prosper's figures and should not be treated as a rate. To get a real number you have to enter their sales process â checked 7 August 2026."},{"question":"Is Prosper AI a good choice?","answer":"For an outpatient group that wants deep specialty-specific voice workflows and has the procurement capacity to run a sales process, yes. Prosper publishes specialty-specific pages for cardiology, dermatology, orthopedics, pediatrics and about a dozen more, plus 80-plus integrations. If your problem is genuinely phone calls and your specialty is on their list, they are a serious option. The friction is that you cannot evaluate cost without a call."},{"question":"What is the difference between a voice agent and a back-office agent?","answer":"A voice agent works the phone: answering patients, calling payers, scheduling, and following up. A back-office agent works the queues that never involve a phone call at all â electronic eligibility checks over the X12 270/271 standard, prior authorization submissions, inbound fax and portal message triage, and denial follow-up. Voice-only vendors handle the first category. Whether you need the second depends on where your staff time actually goes."},{"question":"Should I use voice AI for eligibility verification?","answer":"Only for the part that electronic checks cannot answer. An electronic 270/271 eligibility check costs a fraction of a voice call because nobody waits on hold, and it handles most routine verifications. Voice is worth its cost for complex specialty benefits, carve-outs, and payers that will not respond electronically. A vendor that proposes calling payers by phone for routine eligibility is selling you an expensive answer to a cheap question."}
10,{"question":"Why do healthcare voice AI vendors not publish prices?","answer":"Partly for a real reason and partly for a commercial one. The real reason is that voice cost varies with hold time â a call that waits 20 minutes costs more to run than one answered in two â so a single number misleads. The commercial reason is that quote-gating allows per-customer pricing and requires a sales conversation before a buyer can compare options. In our August 2026 check of 338 healthcare vendors, 83% of the sites we could read gated the product behind a demo or contact form."},{"question":"How should I compare healthcare voice AI vendors?","answer":"Ask each one the same four pricing questions: is hold time billed or only connected time, is there a monthly minimum or annual platform fee on top of usage, what happens to a call that fails or dead-ends, and is there a setup fee. Then ask what share of your staff's time is actually phone work. If most of the burden is faxes, portal messages, and eligibility queues, a voice agent solves the visible problem rather than the expensive one."}]},{"slug":"healthcare-software-demo-gate-research","title":"We Checked 338 Healthcare Vendors. Most Won't Let You Try It.","date":"2026-08-07T14:00:00.000Z","readingTime":"10 min read","excerpt":"We read the primary call to action on every healthcare back-office vendor we could reach. Of the 171 sites we could classify, 142 required a demo or a sales conversation before a practice could see the product. Here's the method, the data, and what we couldn't measure.","category":"Healthcare","segments":["Healthcare"],"tags":["Healthcare","Research","Practice Operations","Vendor Evaluation","Independent Practice"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/demo-gate-research-hero.jpg","heroImageAlt":"A row of many small tan door shapes, most closed with a small padlock, and a few open ones outlined in terracotta showing light through the doorway â vendors that gate the product behind a demo beside the few offering a self-serve trial.","faq":[{"question":"What share of healthcare software vendors require a demo?","answer":"In our August 2026 check of 338 healthcare back-office vendors, we could read the primary call to action on 171 sites. Of those, 142 (83%) required a demo, a sales conversation, or a contact form before showing the product, and 29 (17%) offered a self-serve signup or free trial. That figure describes the 171 sites we could classify, not the whole market â we could not fetch a readable page for the other 167."},{"question":"Why does a demo requirement matter to a small practice?","answer":"It sets a floor on evaluation cost. A demo means scheduling, attending, and usually a follow-up before you learn whether the product does what you need or what it costs. For a five-physician practice with no procurement team, that is real time spent before any information arrives. Self-serve evaluation lets the same practice answer the question in an afternoon."},{"question":"Does a demo requirement mean the product is expensive?","answer":"Not necessarily, and it should not be read that way. Gating can reflect genuinely complex configuration, regulated onboarding, or pricing that legitimately depends on volume and specialty. What it reliably means is that a buyer cannot evaluate the product without entering a sales process, which is a cost regardless of the eventual price."},{"question":"How did you classify each vendor's CTA?","answer":"We fetched each homepage, removed scripts and styles, extracted the text of links and buttons, and matched it against two pattern sets: self-serve language such as sign up, free trial, or create account, and gated language such as get a demo, contact sales, or book a call. Self-serve wins ties, so a site offering both a
10signup and a demo counts as self-serve. Any site we could not fetch or classify was excluded rather than counted."},{"question":"What could you not measure?","answer":"Of 338 vendors, 135 could not be fetched and 32 returned no matching CTA, leaving 51% coverage. Roughly three quarters of the unreachable domains had no DNS record at all, five months after the market map was built. A further 52 sites served no readable HTML without running JavaScript. Those 167 sites might skew in either direction, which is why the finding is stated as a share of the 171 we could read."},{"question":"Which category is most likely to offer self-serve?","answer":"Behavioral health, by a clear margin. Of the behavioral health vendors we could classify, 62% were gated, against 92% in claims management and dental. The likely reason is that behavioral health often sells to solo practitioners, where a sales-led motion does not pay for itself."},{"question":"Can I reproduce this research?","answer":"Yes. The per-site results and the script are published in our repository alongside this post. Re-running it will produce different numbers, because vendor sites change â two vendors in our sample moved from a free trial to a demo requirement in the five months between the original market map and this check."}]},{"slug":"how-to-verify-ai-agent-vendor-claims","title":"How to Verify an AI Agent Vendor's Claims (2026)","date":"2026-08-07T11:00:00.000Z","readingTime":"9 min read","excerpt":"Every healthcare AI vendor publishes a percentage. Almost none publish the denominator. Here's how to read automation claims, what to ask before signing, and how to measure whether an agent actually worked once it's live.","category":"Healthcare","segments":["Healthcare"],"tags":["Healthcare","AI Governance","Vendor Evaluation","Revenue Cycle Management","Measurement"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/verify-agent-claims-hero.jpg","heroImageAlt":"A large tan percentage-style circle on the left sits above an empty outlined box, while on the right a smaller terracotta-outlined circle sits above a box filled with horizontal lines â a claim without a denominator beside a claim with its evidence.","faq":[{"question":"What does a 90% automation claim actually mean?","answer":"On its own, nothing verifiable â the number is meaningless without its denominator. Ninety percent of what? All eligibility checks, or only the ones from payers the vendor already supports? Measured over which period, at which practice, counting a check as automated at what point? The same underlying performance can be presented as 90% or 40% depending on what sits under the line, so the first question about any percentage is what it is a percentage of."},{"question":"What should I ask a healthcare AI vendor before signing?","answer":"Five questions separate a real claim from a demo. What is the denominator behind your headline percentage? What happens to the cases you do not handle â who does that work? What does a failure look like and how will I see it? What is the fully loaded price including minimums, setup fees, and overage? And what will you measure with me after ninety days? A vendor who answers all five in writing is describing a product; one who deflects is describing a demo."},{"question":"How do I measure whether an AI agent actually worked?","answer":"Baseline before you deploy, not after. Record the metric you expect to move â staff hours on the task, denial rate in the relevant category, time to complete â for at least the two weeks before go-live. Then re-measure after a stated interval, typically thirty days for time metrics and a full billing cycle for denial metrics. Without a pre-deployment baseline, any post-deployment number is unfalsifiable."},{"question":"Why do denial improvements take longer to show up than time savings?","answer":"Because claims in flight were filed under the old process. A change to eligibility verification affects claims submitted after go-live, and those take weeks to adjudicate and return. Staff-time savings appear within days; denial-rate changes need a full billing cycle before the comparison means anything. Vendors who report denial improvements in the first two weeks are measuring noise."},{"question":"What is a reasonable automation rate to expect?","answer":"It varies by workflow and should always be a range rather than a single number. Message routing runs about 70 to 80 percent automated because the remainder is genuinely ambiguous. Eligibility verification reaches roughly 90 percent of major payers because coverage is structured data. Any vendor quoting a single precise figure across every practice and payer mix is quoting a demo, not a deployment."},{"question":"Should I trust a case study from a vendor's own customer?","answer":"Trust it as far as it is specific. A useful case study names the practice, the specialty, the starting state, the measurement window, and what did not improve. A weak one gives a percentage and a logo. Ask whether you can speak to the practice directly â vendors with real results usually say yes."}]},{"slug":"inbox-triage-medical-practice-comparison","title":"Your Front Desk Reads Every Message. Most Don't Need a Human.","date":"2026-08-07T10:00:00.000Z","readingTime":"8 min read","excerpt":"Faxes, voicemails, and portal messages arrive in separate queues and s
10omeone reads all of them to find the few that matter. Here's what that costs, how the alternatives compare â hiring, outsourcing, EHR rules, and AI triage â and what automation actually routes.","category":"Healthcare","segments":["Healthcare"],"tags":["Healthcare","Inbox Triage","Practice Operations","Fax Automation","Independent Practice"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/inbox-triage-comparison-hero.jpg","heroImageAlt":"A single tall stack of papers on the left flows through a terracotta sorting point into four small neat stacks on the right, with one stack outlined in terracotta to mark the few messages that need a human.","faq":[{"question":"What is inbox triage in a medical practice?","answer":"Inbox triage is the daily work of reading everything that arrives â faxes, voicemails, patient portal messages, and referral documents â and deciding what each item is, who owns it, and how urgent it is. In most practices it is done manually by front-desk or clinical staff across several separate queues, because faxes, phone systems, and the patient portal rarely share one inbox."},{"question":"How much staff time does message triage consume?","answer":"Two to three hours of daily staff time is typical, and that is the amount practices recover when routing is automated. The cost is not only the reading: each interruption pulls someone away from a patient in front of them, and the switching cost between queues is invisible in any time study."},{"question":"What percentage of practice messages can be routed automatically?","answer":"Seventy to eighty percent of incoming messages can be auto-routed to the right person or workflow. The remaining twenty to thirty percent are genuinely ambiguous, clinically urgent, or unusual enough to need a human decision â which is the correct outcome, not a shortfall."},{"question":"Can my EHR already do this with rules and filters?","answer":"Partly, and it is worth checking before buying anything. EHR rules work well on structured, predictable inputs: a known sender, a consistent subject line, a specific document type. They struggle with a scanned fax that is an image, a voicemail with no text at all, or a portal message where intent has to be read from free text. Rules classify format; triage requires reading content."},{"question":"Is a missed fax really a clinical risk?","answer":"It can be. A hospital discharge summary, a specialist referral, or an abnormal lab result arriving by fax into a shared queue may sit unread for days if nobody claims it. The risk is not that staff are careless â it is that the queue offers no way to tell an urgent document from a marketing flyer without opening both."},{"question":"Does automating triage mean patients talk to a robot?","answer":"No. Triage is about routing, not responding. It reads what arrived, classifies it, and delivers it to the right person or worklist with the context attached. The human who answers the patient is the same human as before, reaching the message sooner with less searching."},{"question":"How is this different from hiring another front-desk person?","answer":"Hiring adds capacity to a process that scales linearly with volume, and it works â it is simply the most expensive option and the slowest to onboard. Automation removes the routine share of the work so existing staff handle exceptions. Practices choosing between them should compare the fully loaded cost of a hire against the share of messages that are genuinely routine, which is most of them."}]},{"slug":"eligibility-verification-cost-comparison-2026","title":"What an Eligibility Check Really Costs (2026 Comparison)","date":"2026-08-07T09:00:00.000Z","readingTime":"10 min read","excerpt":"A manual eligibility verification costs a provider $7.97 per transaction. An electronic one costs $2.18. An automated agent runs about $0.50. Here's the full comparison â portals, clearinghouses, outsourced staff, and AI agents â with the CAQH data behind every figure.","category":"Healthcare","segments":["Healthcare"],"tags":["Healthcare","Eligibility Verification","Revenue Cycle Management","Pricing","Independent Practice","Denial Management"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/eligibility-cost-comparison-hero.jpg","heroImageAlt":"Four insurance cards in a row of increasing size representing rising cost per eligibility check, from a small terracotta-outlined card for an automated check to a large tan card for a manual phone call, with a clock icon above the largest.","faq":[{"question":"How much does an eligibility check cost?","answer":"It depends entirely on how you run it. The 2023 CAQH Index puts the provider cost of a manual eligibility and benefit verification at $7.97 per transaction and a fully electronic one at $2.18. An automated agent that runs the electronic transaction and resolves the exceptions costs about $0.50 per check at Agentman's published rate. The portal itself is free â the cost is the 10 to 15 minutes of staff time it consumes."},{"question":"Why is a payer portal check not free?","answer":"Because the cost is labor, not license fees. A portal check takes 10 to 15 minutes per patient once you count logging in, two-factor authentication, searching, reading the benefits screen, and copying the result into the chart. Across a
10schedule of 30 patients that is several hours of staff time daily. The CAQH Index counts that labor, which is why it prices a manual verification at $7.97 to the provider even though no one sends an invoice."},{"question":"What is the difference between a manual, partial, and electronic eligibility check?","answer":"CAQH classifies them by how the transaction runs. Manual means phone, fax, or a payer web portal â a human does the work. Partial means a mix, typically an electronic query whose gaps get resolved by phone. Electronic means a standard X12 270/271 transaction with no human in the loop. The 2023 CAQH Index prices these to the provider at $7.97, $4.07, and $2.18 respectively."},{"question":"Is an AI eligibility agent cheaper than a clearinghouse?","answer":"They do different amounts of the job. A clearinghouse charges per transaction for running the 270/271 and returns whatever the payer sends back, including incomplete responses. An agent runs the same transaction and then resolves the exceptions â bad member IDs, missing plan details, coverage that terminated â which is the part that otherwise lands back on staff. Compare cost per completed verification, not cost per transaction."},{"question":"How many denials come from eligibility problems?","answer":"Eligibility and coverage errors consistently rank among the top denial causes across specialty practices. In Agentman customer practices the category is large enough to be worth automating first. Valley Diabetes \u0026 Obesity saw a 65% reduction in eligibility and prior-authorization-related denials after automating verification."},{"question":"Does automating eligibility verification replace front-desk staff?","answer":"No. It removes a specific repetitive task that consumes one to two hours per provider per day. The front-desk role shifts toward patient communication, exception handling, and the complex cases that require judgment. Practices that automate eligibility typically redeploy that time rather than reduce headcount."},{"question":"How far in advance should eligibility be verified?","answer":"Five days before the appointment, not the day before. A five-day window leaves time to reach a patient whose coverage is invalid, update their insurance, or rebook the slot. Day-before verification leaves no recovery window: the slot goes empty or the visit becomes uncompensated care."}]},{"slug":"healthcare-voice-ai-pricing-comparison","title":"What Healthcare Voice AI Actually Costs (2026 Comparison)","date":"2026-08-06T12:00:00.000Z","readingTime":"10 min read","excerpt":"Every major healthcare voice AI vendor gates pricing behind a custom quote. We compared what each one publishes â and published our own rate, $10 for each hour of call time â so you can budget before you book a demo.","category":"Healthcare","segments":["Healthcare"],"tags":["Healthcare","Voice AI","Revenue Cycle Management","Pricing","Vendor Comparison","Independent Practice"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/healthcare-voice-ai-pricing-hero.jpg","heroImageAlt":"Five telephone handsets in a row. Four tan handsets each have a speech bubble containing a closed padlock; the fifth, outlined in terracotta, has a speech bubble showing visible price lines â the four healthcare voice AI vendors that quote privately beside the one that publishes its rate.","faq":[{"question":"How much does healthcare voice AI cost per call?","answer":"It depends on call length, because payer calls are billed by time spent on the phone. At Agentman's published rate of $10 per hour of call time, a 12-minute payer call costs about $2.00 and a 3-minute patient scheduling call costs about $0.50. Most competing vendors do not publish a rate, so a per-call figure for them cannot be calculated without a quote."},{"question":"Why do most healthcare voice AI vendors hide their pricing?","answer":"Two reasons, one legitimate and one commercial. The legitimate one is that voice cost genuinely varies with hold time â a call that waits 20 minutes on hold costs more to run than one answered in two, so vend
10ors avoid quoting a single number. The commercial one is that quote-gating enables per-customer pricing and requires a sales conversation before a buyer can compare options."},{"question":"Is voice AI cheaper than an electronic eligibility check?","answer":"No, and it should not be used for the same job. An electronic 270/271 eligibility check costs a fraction of a voice call because no one waits on hold. Voice is for the work portals cannot do: complex specialty benefits, carve-outs, claim follow-up, and any payer that will not answer electronically. Use electronic checks for the routine 60 to 80 percent and voice for the remainder."},{"question":"What should I ask a voice AI vendor about pricing?","answer":"Ask four questions. Is hold time billed or only connected time? Is there a monthly minimum or annual platform fee on top of usage? What is the rate for a call that fails or reaches a dead end? And is there a setup or onboarding fee? The published rate matters far less than what the contract counts as billable."},{"question":"Does a voice agent replace the rest of revenue cycle automation?","answer":"No. Voice handles work that requires a phone call. Eligibility verification, prior authorization submission, and inbox triage are not phone problems, and a voice-only vendor cannot do them. A practice evaluating voice should check whether the vendor also covers the non-voice work, or whether voice is a single point solution alongside other tools."},{"question":"Do healthcare voice AI vendors require a minimum spend commitment?","answer":"Some do, and it is rarely disclosed publicly. In July 2026, a sales inquiry we sent to Assort Health came back stating that a $75,000 annual minimum commitment was required before an initial call could be scheduled, and that the minimum reflected a floor of five providers. That is about $15,000 per provider per year, or roughly $1,250 per provider per month, with a five-provider floor underneath it. Both halves matter: the rate tells you what the service costs, and the floor tells you whether a practice your size is in scope at all. A minimum stated in dollars reads as a price; stated in providers, it can mean you are not the customer. Ask any vendor whether a minimum applies to a practice your size, and ask what unit that minimum is counted in."},{"question":"How do I compare voice AI vendors when only one publishes pricing?","answer":"Normalize everything to cost per completed task, not per minute or per seat. Ask each vendor to quote the same concrete scenario â for example 200 benefit verification calls per month at your typical hold times â and require the quote to include platform fees, minimums, and overage rates. That converts incomparable pricing models into one number you can compare."}]},{"slug":"share-the-skill-not-the-secret-use-only","title":"Share the Skill, Not the Secret: Use-Only Access and the New Economics of Expertise","date":"2026-07-06T18:00:00.000Z","readingTime":"9 min read","excerpt":"A shared document is a copied document. A use-only skill is runnable but not readable â the body is redacted at the API layer, and every use is logged. That one distinction changes what experts are willing to share, and with whom.","category":"Governance","segments":["Governance","Agent Skills"],"tags":["Governance","Agent Skills","Claude Skills","Use-Only Access","IP Protection","Access Control"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/use-only-share-the-skill-hero.jpg","heroImageAlt":"Use-only access â a padlocked skill whose instructions stay sealed while its output flows to a teammate who runs it, with a closed-eye glyph signaling not-readable","faq":[{"question":"Can a use-only user really not see the instructions?","answer":"No, they can't â and not because a button is hidden. The skill body is redacted at the API layer, so the instructions are never sent to the user's client at all. There's nothing in the browser, the network response, or developer tools to reveal. The skill executes server-side and only its output returns."},{"question":"Does use-only work with people outside my organization?","answer":"Yes. That's one of its main purposes â a client, co-investor, or outside agency can run your skill without it leaving your control. For a hard confidentiality boundary, run a separate workspace per client, each with its own permissions and logs, and share the skill use-only within it."},{"question":"Is there a record of who used a use-only skill?","answer":"Yes. Every invocation is logged with the identity of who ran it and when, visible in the per-skill activity log, and share/use events roll up into the workspace access log. Use-only gives you confidentiality of the contents and visibility of the usage."},{"question":"What's the difference between use-only and just not sharing edit access?","answer":"Read access still lets someone open and copy the instructions â they just can't change them. Use-only removes read entirely: the person can run the skill but never see what's inside. If the skill's value is its contents, that's the difference between protected and exposed."}]},{"slug":"skill-building-best-practices-draft-publish-version","title":"Draft, Publish, Version: Best Practices for Building Agent Skills That Last","date":"2026-07-06T17:00:00.000Z","readingTime":"8 min read","excerpt":"The best skill builders treat drafts as a live workshop â a draft skill already works in Claude, so you test it in real conversations, then publish only when you're satisfied. Here's the draft â publish â version workflow, and the habits that make skills durable.","category":"Agent Skills","segments":["Agent Skills","Claude Skills"],"tags":["Agent Skills","Claude Skills","SKILL.md","Version Control","Best Practices","How-To"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/skill-lifecycle-draft-publish-version-hero.jpg","heroImageAlt":"The draft, publish, version workflow for building durable Agent Skills â a draft works in Claude, publishing freezes a stable version, and version history keeps the record","faq":[{"question":"Can I use a skill in Claude before I publish it?","answer":"Yes. A draft skill is already usable in Claude once your library is connected â that's the whole point of the draft-first workflow. You test the skill in real conversations, refine it, and publish only when you're satisfied. Publishing isn't what makes a skill work; it's what freezes a version as the stable, shareable one."},{"question":"What does publishing actually do, then?","answer":"Two things: it freezes the current version so it's immutable (which is what makes version history trustworthy), and it marks that version as the stable one your teammates and agents should rely on. It does not change who can see the skill â visibility is a separate sharing decision."},{"question":"Will I run out of published-skill slots while experimenting?","answer":"No. Drafts are unlimited on every plan, and re-publishing a new version of an already-published skill doesn't use another slot â only distinct published skills count. Experiment freely."},{"question":"What happens if I edit a skill that's already published?","answer":"Editing a published skill automatically opens a new draft, and the published version keeps running untouched until you publish again. You can't break the live version by experimenting â and you can diff the new draft against the live one before you commit."}]},{"slug":"agent-skills-relaunch-team-layer-claude-skills","title":"Agent Skills, Relaunched: The Team Layer for Your Claude Skills","date":"2026-07-06T16:00:00.000Z","readingTime":"4 min read","excerpt":"Six weeks of shipped work in one release: Google-Docs-style sharing, use-only access that protects your IP, version history 2.0 with one-click restore, governance logs, batch import for the skills you built in Claude, and new pricing that starts free.","category":"Agent Skills","segments":["Agent Skills"],"tags":["Agent Skills","Claude Skills","Product Update","Announcement","Blog"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/agent-skills-relaunch-hero-v2.jpg","heroImageAlt":"The team layer for Claude Skills â one shared skill at the center, distributed to a team, wrapped in access control, version history, and activity logging","faq":null},{"slug":"claude-skills-vs-agent-skills","title":"Claude Skills vs Agent Skills: Same Format, Different Layer (A Practical Guide)","date":"2026-07-06T16:00:00.000Z","readingTime":"10 min read","excerpt":"Claude Skills and Agent Skills are the same open SKILL.md standard â the difference is where they live and who governs them. An honest comparison, including what Claude's native org sharing already does well and the four gaps a governed library fills.","category":"Agent Skills","segments":["Agent Skills","Claude Skills"],"tags":["Agent Skills","Claude Skills","SKILL.md","AI Governance","Comparison","Team Collaboration"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/claude-vs-agent-skills-hero.jpg","heroImageAlt":"Claude Skills vs Agent Skills â the same SKILL.md document, one plain and one wrapped in a governance layer of use-only access, version history, and activity logging","faq":[{"question":"Are Claude Skills and Agent Skills different formats?","answer":"No. They're the same open SKILL.md standard Anthropic released in October 2025 â a folder with a markdown instruction file, optional scripts, and reference files. A skill written in Claude imports into an Agent Skills library without conversion, and runs unchanged."},{"question":"Can't Claude already share skills with my team?","answer":"Yes â since December 18, 2025, Claude Team and Enterprise plans include native organization sharing: owner provisioning, peer-to-peer sharing, and an internal directory. It handles distribution well. What it doesn't include is running a skill without reveal
10ing its contents, a record of who used a skill, in-product version history with restore, or an approval workflow â the governance layer Anthropic's own enterprise guide suggests teams build themselves."},{"question":"Does using an Agent Skills library mean leaving Claude?","answer":"No. You connect the library to Claude over MCP and keep working in Claude exactly as before. The library is where the skills live and how they're governed; Claude is still where you run them â along with Claude Code, ChatGPT, and Cursor from the same library."},{"question":"What's the single biggest difference?","answer":"Whether \"shared\" also means \"readable and copyable.\" In native sharing it does. A governed library adds use-only access, so a teammate â or an outside firm â can run your skill without ever seeing what's inside it, with every use logged."}]},{"slug":"why-your-pe-firms-best-analyst-is-a-skills-library","title":"Why Your PE Firm's Best Analyst Is a Skills Library","date":"2026-06-26T21:52:43.778Z","readingTime":"11 min read","excerpt":"A private equity skills library beats individual analysts on consistency, speed, and institutional memory â automating CIM extraction, comps, IC memos, and DDQs so your analysts spend their hours on judgment instead of data plumbing.","category":"PE","segments":["PE"],"tags":["Private Equity","Agent Skills","Agentic AI","Due Diligence","Thought Leadership"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/pe-skills-library-hero.jpg","heroImageAlt":"Flat-design illustration of a private equity skills library: deal documents (CIM, DDQ, comps) flow through a grid of labeled AI agent-skill cards into an analyst freed for judgment.","faq":[{"question":"What is a skills library in private equity?","answer":"A skills library is a governed, versioned set of AI agents that perform a PE firm's repeatable analytical work â extracting CIM data, building comps, drafting IC memos, and answering diligence questionnaires â using the firm's own criteria. It is owned infrastructure, so the firm's standards stay consistent across deals and persist even as analysts rotate."},{"question":"Will an AI skills library replace PE analysts?","answer":"No. A skills library automates the data extraction and formatting that consume most of an analyst's hours, freeing analysts for judgment, relationships, and deal strategy. The aim is leverage â one analyst plus a skills library producing the output of several â not headcount reduction. The strategic decisions remain human."},{"question":"How long does it take to build a PE skills library?","answer":"A focused pilot on a single workflow, such as CIM extraction, can be live in roughly 2â4 weeks. Full integration with a mid-sized firm's existing stack â data rooms, CRM, reporting â typically takes 3â6 months. The proven path is to start narrow, prove ROI on one skill, then expand."}]},{"slug":"agent-skills-ecosystem-report-2026","title":"The Agent Skills Ecosystem in 2026: Who's Building, What's Working, and What's Next","date":"2026-06-25T00:00:00.000Z","readingTime":"13 min read","excerpt":"A 2026 landscape map of the agent skills ecosystem â the adopters, directories, fastest-growing categories, the first peer-reviewed quality benchmark, and where the value is heading in H2.","category":"Thought Leadership","segments":["Thought Leadership"],"tags":["Agent Skills","Agent Skills Ecosystem","AI Agents","Thought Leadership","SKILL.md","AEO"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/agent-skills-ecosystem-report-2026-hero.jpg","heroImageAlt":"Flat geometric illustration of the agent skills workflow: SKILL.md files organized into a curated, version-controlled library, assembled into an AI agent that runs across multiple AI tools.","faq":[{"question":"What is an agent skill?","answer":"An agent skill is a folder containing a SKILL.md file plus optional scripts and resources that teach an AI agent how to perform a specific task consistently. The agent reads the skill's name and description to decide when to use it, then loads the full instructions only when a task matches."},{"question":"Is the Agent Skills standard open?","answer":"Yes. Anthropic published the Agent Skills specification as an open standard at agentskills.io on December 18, 2025, and stewards it through the Agentic AI Foundation. A skill written to the spec runs across compatible platforms without modification."},{"question":"Which AI tools support agent skills?","answer":"As of June 2026, roughly 40 products on the agentskills.io showcase support the standard, including Claude, OpenAI Codex, GitHub Copilot, VS Code, Cursor, Gemini CLI, Goose, OpenCode, Databricks Genie Code, and Snowflake Cortex Code."},{"question":"Are community skills safe to install?","answer":"Not automatically. A 2026 audit found prompt injection in 3
106% of tested skills, and skills can execute bundled scripts. Review the source of any community skill before installing it, exactly as you would an open-source dependency."},{"question":"Do bigger skill catalogs mean better skills?","answer":"No. SkillsBench found the average public skill scores 6.2 out of 12, and only top-quartile skills meaningfully improved agent performance. Curation and review matter more than catalog size."}]},{"slug":"pe-intelligence-stack-agent-skills","title":"40 Skills, 8 Layers, 6 Solutions: The PE Intelligence Stack","date":"2026-06-22T20:10:06.609Z","readingTime":"13 min read","excerpt":"Agentman's full-stack PE platform â 40 agent skills across 8 capability layers, from data intelligence to valuation, governance, and orchestration â composing into 6 solutions. Sixteen are live in the library today.","category":"PE","segments":["PE"],"tags":["Private Equity","Agent Skills","Product","Platform","Deal Intelligence"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/pe-intelligence-stack-hero.jpg","heroImageAlt":"Flat-design illustration of a private-equity deal pipeline: a CIM and deal documents flow through an AI agent hub into an IC memo, a performance dashboard with a portfolio-company building, and an LP report with a compliance badge.","faq":[{"question":"What is the PE Intelligence Stack?","answer":"The PE Intelligence Stack is Agentman's full-stack agent skills platform for private equity: 40 agent skills across 8 capability layers that compose into 6 packaged solutions on the Agent OS platform. Sixteen skills are live in the library today; the rest are platform-native capabilities or planned releases. Each skill is a versioned, auditable procedure rather than a prompt."},{"question":"What can AI agents do in private equity?","answer":"AI agents in private equity automate document-heavy, multi-step work: screening targets against an investment thesis, extracting metrics from CIMs and contracts, building LBO and DCF models, drafting IC memos, monitoring portfolio KPIs, and generating LP reports. Each task runs as a structured skill with a full citation trail to its source."},{"question":"How is an agent skill different from a ChatGPT prompt?","answer":"A prompt is a single instruction with variable output. An agent skill is a versioned procedure that specifies exactly what to extract, how to flag it against a firm's taxonomy, and where to place it â producing consistent, citable, review-ready output every time, with orchestration and governance handled by the platform."},{"question":"Is AI in private equity compliant and auditable?","answer":"It can be, when governance is a layer of the stack. Agentman's Governance \u0026 Compliance layer provides compliance validation against SEC and LPA requirements and EDCI-aligned ESG reporting, while the Agent OS platform supplies audit trails, citation, access control, and data lineage â addressing both LP due-diligence questions and EU AI Act requirements taking effect in August 2026."},{"question":"How many AI skills does a PE firm need to start?","answer":"One. Most firms clone a single high-value skill â a CIM Analyzer or Target Screener, both live today â configure it to their criteria, and run it in days. Value compounds as additional skills are chained into end-to-end workflows across the full 8-layer stack."}]},{"slug":"independent-medical-practice-back-office-pain-points-ai-agents","title":"The 8 Back-Office Pain Points Crushing Independent Medical Practices â and the AI Agents That Fix Them","date":"2026-06-15T19:58:14.466Z","readingTime":"14 min read","excerpt":"Independent medical practices lose hours and revenue to eligibility, prior auth, denials, and unmanaged inboxes. Here are the 8 biggest back-office pain points and the Agentman agents built to solve each â two live today, six on the 2026 roadmap.","category":"Healthcare","segments":["Healthcare"],"tags":["Healthcare","Revenue Cycle Management","AI Agents","Prior Authorization","Eligibility Verification","Claim Denials","Medical Practice"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/independent-practice-pain-points-8-agents-hero.jpg","heroImageAlt":"Flat geometric illustration titled \\\\\\\"8 Pain Points Crushing Independent Medical Practices\\\\\\\" showing eight agent cards representing a medical practice back office","faq":[{"question":"How many agents does Agentman have, and which are live?","answer":"Agentman has eight agents covering the medical practice back office: eligibility verification, inbox triage, prior authorization, denial discovery, claims coding, claims submission, patient communications, and Rx refills. Two are deployed in live practices today â eligibility verification and inbox triage. The remaining six are in active development, scheduled across 202
106."},{"question":"How much does Agentman's eligibility verification agent cost per check?","answer":"Agentman's eligibility verification agent runs real-time checks across more than 2,700 payers at roughly $0.50 per check. That compares against a manual eligibility-check cost Agentman benchmarks at approximately $7.97, reflecting the staff time and phone-and-portal work that manual verification requires."},{"question":"Does Agentman require EHR integration to start?","answer":"No. Agentman's deployed agents run the back office without requiring EHR integration to begin, which lowers the barrier for small independent practices. The Rx refill agent on the roadmap is designed to integrate with the EHR for auto-approval workflows."},{"question":"Is Agentman HIPAA compliant?","answer":"Yes. Agentman is HIPAA compliant and SOC 2 ready, with full data lineage and governance â audit trails, access controls, and traceability from patient record to claim submission â built into the platform architecture."}]},{"slug":"agent-skills-compound-enterprise-intelligence","title":"From 1 Skill to 100: How Agent Skills Compound Into Enterprise Intelligence","date":"2026-06-12T19:00:00.000Z","readingTime":"9 min read","excerpt":"Skills don't just add value â they compound. From a single $0.50 eligibility check to chained claim-to-appeal workflows to an 87+ skill library where every new skill upgrades the rest: here's the progression from one skill to enterprise intelligence, with worked examples from healthcare RCM and private equity.","category":"Agent Skills","segments":["Agent Skills"],"tags":["Agent Skills","Cross-Vertical","Education","Use Cases","Network Effects","AI Agents","Revenue Cycle Management","Private Equity"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/agent-skills-compound-enterprise-intelligence-hero.jpg","heroImageAlt":"Flat geometric infographic showing three labeled stages â 1 skill (eligibility check), skills chained into a workflow, and an 87+ skill network â above a rising compounding curve in Agentman terracotta and charcoal.","faq":[{"question":"How many agent skills do I need before compounding starts?","answer":"Compounding typically becomes visible at three to four skills, the point where one skill's output feeds another. Before that, skills deliver standalone time savings. After that, each addition upgrades existing workflows, and a 10-skill library already supports 45 possible two-skill combinations."},{"question":"What's the difference between an agent skill and a prompt?","answer":"A prompt is a one-time instruction that produces a single answer. An agent skill is a versioned, testable package of instructions, domain knowledge, and tool access that produces consistent results every run. Repeatability is what allows skills to be chained into workflows and shared across a team."},{"question":"Can I use pre-built skills or do I have to build my own?","answer":"You can start with pre-built skills. The myAgentSkills.ai library offers 87+ production-ready skills to clone and customize on day one. Most teams clone a library skill, configure it for their payers, criteria, or style, and have it running in days rather than building from scratch over weeks."},{"question":"Which skill should a medical practice deploy first?","answer":"Start with eligibility verification. It is high-volume, binary in outcome, and easy to measure: Agentman's eligibility verification agent runs at $0.50 per check versus the $7.97 CAQH Index benchmark. Once it's trusted, chain in prior authorization, claim tracking, and denial management."}]},{"slug":"agent-skills-transforming-healthcare-claims-appeals-clinical-notes","title":"How Are Agent Skills Transforming Healthcare? From Claims Appeals to Clinical Notes","date":"2026-06-12T00:00:00.000Z","readingTime":"11 min read","excerpt":"Agent Skills package expert RCM knowledge â denial playbooks, coding audit methodology, PA criteria â into files AI agents execute on demand. Here's how independent specialty practices use them to automate claims appeals, billing compliance, prior authorization, and clinical documentation, with 2025â2026 AMA, KFF, and HFMA data.","category":"Healthcare","segments":["Healthcare"],"tags":["Agentic Healthcare","Agent Skills","Use Cases","Denial Management","Prior Authorization","Medical Billing","Revenue Cycle Management","Clinical Documentation Improvement","Independent Practice","HIPAA"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/agent-skills-healthcare-claims-appeals-hero.jpg","heroImageAlt":"Flat geometric illustration showing denied claims documents flowing through an Agent Skills node with modular skill cards into clean claims and clinical notes with an approval check mark.","faq":[{"question":"What is the difference between an AI agent and an Agent Skill?","answer":"An AI agent is the system that performs work â reading documents, making API calls, drafting outputs. An Agent Skill is the packaged expertise the agent loads for a specific task: procedures, decision rules, templates, and reference data. One agent can load many skills; one skill can be shared across many agents and organizations."},{"question":"Can a small medical practice use Agent Skills without an IT team?","answer":"Yes. Skills are designed to be loaded by an agent platform, not installed like softw
10are. A practice manager describes the task â \"analyze these denials\" or \"check PA criteria for this medication\" â and the agent selects and applies the matching skill. Setup effort concentrates in connecting data sources (clearinghouse reports, EOBs/ERAs) once."},{"question":"Do Agent Skills replace billers and coders?","answer":"No. Skills shift staff from production work to exception handling and oversight. The agent screens 100% of claims, drafts 100% of routine appeals, and routes the ambiguous or high-stakes cases to humans. Credentialed coders still own final coding decisions, and clinicians remain the authors of clinical documentation."},{"question":"How do skills stay current with changing payer rules?","answer":"Skills are versioned files, so payer rule changes become skill updates â edited once, applied everywhere the skill runs. This is a structural advantage over tribal knowledge, which updates only when an individual staff member happens to learn about a change."}]},{"slug":"build-your-first-agent-skill-skillmd-anatomy","title":"How Do You Build Your First Agent Skill? A Complete SKILL.md Anatomy Guide","date":"2026-06-09T17:11:51.897Z","readingTime":"16 min read","excerpt":"A complete SKILL.md tutorial (the format behind Claude Skills): the two required frontmatter fields, the optional ones, the three-tier progressive disclosure model, when to add scripts and references, and three ways to build your first agent skill â by hand, by chatting with Claude, or with Agentman's no-code builder.","category":"Agent Skills","segments":["Agent Skills"],"tags":["Agent Skills","SKILL.md","Tutorial","Progressive Disclosure","No-Code","AI Agents","Education"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/build-your-first-agent-skill-hero.jpg","heroImageAlt":"Flat illustration showing three paths to creating an agent skill â write it yourself, chat with Claude, or use a no-code builder â all converging into a single SKILL.md file with name and description frontmatter.","faq":[{"question":"What are the required fields in a SKILL.md file?","answer":"A SKILL.md file requires exactly two YAML frontmatter fields: name and description. The name must be lowercase with hyphens, stay under 64 characters, and match the parent folder name. The description must stay under 1,024 characters and explain both what the skill does and when to use it. All other fields are optional."},{"question":"How is a skill different from just writing a long prompt?","answer":"A skill uses progressive disclosure, so only its name and description load at startup and the full instructions load only when a task matches. A long prompt consumes context every time, whether or not it is relevant. Skills also persist as reusable files and run across any agent that supports the open standard, while a prompt lives only in one conversation."},{"question":"How long should the SKILL.md body be?","answer":"The recommended ceiling for the SKILL.md body is about 5,000 tokens, because the agent loads the entire body when it activates the skill. Content beyond that should move into the references/ directory, which the agent loads only when the instructions point to it. Keeping the body lean protects the context window and speeds activation."},{"question":"Do skills built for one agent work on others?","answer":"Yes, if they follow the open specification at agentskills.io. The standard defines the file structure, frontmatter schema, and loading behavior, but leaves routing to each runtime. A spec-compliant skill written for Claude Code generally runs on VS Code Copilot, Cursor, and OpenAI Codex, though platform-specific fields may be ignored by runtimes that do not use them."},{"question":"Can you create a skill without writing any code?","answer":"Yes. In Claude.ai or Cowork, ask Claude to use its built-in skill-creator, which drafts the SKILL.md from a plain-language description and packages it for you. Agentman's myAgentSkills.ai builder goes a step further by hosting the skill and sharing it across your team over MCP, so there is no folder to set up or re-upload. Code execution must be enabled to run any bundled scripts."}]},{"slug":"three-layer-intelligence-model","title":"The Three-Layer Intelligence Model: Why Your AI Gets Smarter With Every Transaction","date":"2026-06-05T21:55:26.643Z","readingTime":"12 min read","excerpt":"A layered model â industry intelligence at the base, firm-specific knowledge in the middle, and transaction-level intelligence on top, each codified as a skill â beats monolithic AI and compounds with every transaction.","category":"Platform","segments":["Platform"],"tags":["Agent Skills","Agentic AI","AI Architecture","Healthcare AI","Private Equity","Revenue Cycle Management"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/three-layer-intelligence-model-hero.jpg","heroImageAlt":"Diagram of the three-layer intelligence model: Industry Intelligence at the base, Firm/Practice-Specific Knowledge in the middle, and Transaction-Level Intelligence on top, each codified as a skill, with value compounding upward into productivity and revenue.","faq":[{"question":"What is the three-layer intelligence model in AI?","answer":"The three-layer intelligence model organizes AI into three independent tiers: industry intelligence as the shared base, firm- or practice-specific knowledge in the middle, and transaction-level intelligence on top. Each tier is built and maintained as a separate skill, and the top layer learns from every transaction it processes."},{"question":"Why is layered AI better than a single large model?","answer":"Layered AI is better because specialized components can be built, tested, and improved independently, and research shows compound systems outperform single monolithic models on real-world tasks (Zaharia et al., 2024). Layering also lets a new team inherit a ready-made base layer instead of training one from scratch."},{"question":"How does the three-layer model reduce claim denials?","answer":"The model reduces denials by combining shared payer-policy rules at the base layer with a practice's specific workflows in the middle layer to handle each claim and denial at the top layer. Because initial denials reached 11.8% in 2024 and many are never resubmitted, an agent that recognizes recurring payer patterns recovers revenue that would otherwise be lost (Experian Health, 2025)."}]},{"slug":"inbox-triage-agent-medical-practice-10-months","title":"Newsletter #6: The Fax We Missed, the Voicemail That Caught It: Ten Months Inside Our Inbox","date":"2026-06-03T21:01:53.170Z","readingTime":"13 min read","excerpt":"We deployed an inbox triage agent at Valley Diabetes \u0026 Obesity to clear a 47-fax morning pile. Ten months later it auto-routes 80% of inbound fax, email, and voicemail and has recaptured $50K+ â but the real story is the 20% it couldn't classify, where device recalls, self
10-generated prior auths, and hospital hand-backs were quietly slipping through.","category":"Healthcare","segments":["Healthcare","Newsletter"],"tags":["Newsletter","The Agentic Practice","Agentic Healthcare","Inbox Triage","Referral Management","Prior Authorization","Eligibility Verification","Revenue Cycle Management","Independent Practice","Practice Operations","Fax Automation"],"author":{"name":"Sachin Gangupantula","avatar":"/images/sachin-gangupantula.jpeg"},"heroImage":"/images/blog/News6_image1.png","heroImageAlt":"The Agentic Practice Issue 6 cover: The Fax We Missed, the Voicemail That Caught It â ten months inside a specialty practice's inbox","faq":null},{"slug":"plain-english-skill-building-for-domain-experts","title":"You Don't Need AI Engineers to Build AI Agents: A Plain-English Guide for Domain Experts","date":"2026-06-01T19:40:52.696Z","readingTime":"9 min read","excerpt":"Domain experts can build production AI agent skills without code or AI engineers. Describe the behavior, attach examples, generate the skill. Here's how.","category":"Agent Skills","segments":["Agent Skills"],"tags":["Agent Skills","Cross-Vertical","Domain Experts","Agentic AI"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/plain-english-skill-building-hero.jpg","heroImageAlt":"Split illustration: on the left, a domain expert sidelined behind an AI barrier filing requirement tickets; on the right, the same expert describing the behavior in plain English, leading to a productivity boost.","faq":[{"question":"Do I need to know how to code to build an agent skill with Agentman?","answer":"No. Plain-English skill building requires no coding, prompt engineering, or machine learning expertise. You describe the behavior in natural language, attach reference documents, and the platform generates the skill. The expertise you need is in your own domain, not in AI."},{"question":"What is an agent skill?","answer":"An agent skill is a reusable unit of agent behavior â a specific task the agent performs consistently, such as verifying insurance eligibility or drafting an investment committee memo. A skill defines the inputs, steps, and outputs for that task so the agent runs it the same way every time."},{"question":"How is plain-English skill building different from hiring an AI engineer?","answer":"An AI engineer understands models but must learn your domain before building anything useful, creating a slow, lossy handoff. Plain-English skill building lets the domain expert build the skill directly, so the expert's judgment is captured without a translation layer â faster, and with higher fidelity to how the work is actually done."},{"question":"What kinds of reference documents should I attach?","answer":"Attach completed examples that represent your standard: sample prior authorizations, claim examples, memo templates, CIMs, LP report formats, or finished work products. The quality of a generated skill depends directly on the quality and relevance of the references you provide."},{"question":"What happens to a skill when an employee leaves?","answer":"The skill stays. Because plain-English skill building encodes repeatable expertise into a reusable asset, institutional knowledge remains with the organization rather than leaving with the individual. New hires inherit a working system instead of starting from scratch."}]},{"slug":"build-ai-agent-skills-no-code","title":"You Don't Need to Code to Build AI Agent Skills (Here's Proof)","date":"2026-05-26T21:38:12.490Z","readingTime":"10 min read","excerpt":"You don't need to code to build AI agent skills. Here's the plain-English, no-code workflow to take a non-developer from idea to a deployed agent skill in under 10 minutes â plus how the same approach powers Medman's healthcare agents.","category":"Platform","segments":["Platform"],"tags":["Agent Skills","Platform","No-Code","AI Agents","Healthcare AI"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/build-ai-agent-skills-no-code-hero.jpg","heroImageAlt":"Flat geometric illustration of a smiling woman at a desk thinking \\\\\\\"Agent Skills,\\\\\\\" with output flowing into a glowing lightbulb-brain icon with electrodes","faq":[{"question":"Do you need to know how to code to build an AI agent skill?","answer":"No. An AI agent skill is
10written in plain English as a structured set of instructions, not in a programming language. The skill describes what task to do and how, and the AI agent interprets and executes it. The core requirement is clear thinking about the task, not coding ability."},{"question":"How long does it take to build a no-code AI skill?","answer":"A well-scoped skill can go from idea to deployed in under 10 minutes using Agentman's no-code builder. The time depends on how clearly defined the task is, not on technical complexity. Simple, single-purpose skills are fastest; refining a skill against real examples adds a few minutes but improves reliability."},{"question":"What makes an AI agent skill instruction effective?","answer":"Effective instructions are specific, scoped to one action per step, declarative, and explicit about edge cases. Vague instructions like \"handle my emails\" underperform specific ones that define what to do, in what order, and what to skip. Over-stuffing or repeating instructions reduces reliability."},{"question":"Can non-technical healthcare staff build their own automation skills?","answer":"Yes. Practice managers and billing staff can author skills that capture their own workflows â like eligibility verification or denial follow-up â in plain English. Clinical or patient-facing skills require human and clinician review before deployment, but the authoring itself requires no code."}]},{"slug":"bridging-healthcare-gaps-agentic-healthcare","title":"Bridging Healthcare Gaps (YouTube): Why Agentic Healthcare Matters â Real-World Use Cases, Risks, and Opportunities","date":"2026-05-22T18:18:17.647Z","readingTime":"15 min read","excerpt":"Agentman's Prasad Thammineni and Sachin Gangupantula joined Tyler Lo on the Bridging Healthcare Gaps podcast to explain why agentic AI matters for independent specialty practices â from eligibility verification that protects $50Kâ$80K per physician per year to the agent-skills flywheel, plus the governance and risk questions that come with clinical AI.","category":"Newsroom","segments":["Newsroom"],"tags":["Newsroom","Bridging Healthcare Gaps","Agentic Healthcare","Eligibility Verification","Revenue Cycle Management","Independent Practice","Agent Skills","The Agentic Practice"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/newsroom/bridging-healthcare-gaps-youtube.jpg","heroImageAlt":"Prasad Thammineni and Sachin Gangupantula of Agentman (Chain of Agents) featured on the Bridging Healthcare Gaps podcast hosted by Tyler Lo.","faq":[{"question":"What is an eligibility verification agent?","answer":"An eligibility verification agent is software that automatically confirms a patient's insurance coverage, plan status, and provider assignment before an appointment. It connects to a practice's management system, checks coverage days in advance, and flags patients whose insurance is invalid so staff can resolve the issue before the visit."},{"question":"How much does eligibility verification automation save a medical practice?","answer":"An eligibility agent saves an independent practice an estimated $50,000â$80,000 per physician per year by preventing denials from ineligible visits, plus eight or more hours of staff time per week. Agentman reports these figures from live deployment at a specialty practice over a three-month period."},{"question":"How long does it take to set up an eligibility verification agent?","answer":"A practice can sign up, connect its practice management system, and run its first eligibility check in about five minutes, with measurable time savings in the first hour. Agents that require clinical training, such as prior authorization, take longer to configure for specific procedures and payers."},{"question":"Does an eligibility verification agent work with every insurance payer?","answer":"No. Some local or smaller payers do not support programmatic eligibility checks. In Agentman's live deployment, the agent automated roughly 90% of the practice's highest-volume payers, with the remainder still requiring manual verification."}
10,{"question":"How is Agentman's eligibility verification agent priced?","answer":"Agentman prices its eligibility agent at $150 per physician per month, or about $1,800 per year. A usage-based option is also available at roughly the same effective rate for a practice seeing about 30 patients per day."},{"question":"What other administrative work can an agentic back office automate?","answer":"Beyond eligibility, an agentic back office automates inbox and fax triage, prior authorization, denial management, referrals, and patient intake. A single-physician practice receives an estimated 80â90 voicemails and 100â120 faxes daily, and a universal inbox agent triages this volume so staff can prioritize urgent items like hospital discharge notices."},{"question":"What is an agent skill in healthcare automation?","answer":"An agent skill is a set of plain-English instructions describing how a human performs a workflow â such as assembling a prior authorization for a specific drug â that an agent loads on demand. Skills are reusable across similar practices, so each new practice in a specialty configures faster than the last."}]},{"slug":"ai-reinvention-intelligence-architect","title":"Podcast by Camber Creek: The Man Who Helped Salesforce Prepare for the AI Era Has Advice for the Rest of Us","date":"2026-05-20T20:29:24.042Z","readingTime":"12 min read","excerpt":"AI agents won't replace knowledge workers â they'll reinvent them. In a conversation on Camber Creek's Catalyst podcast, Agentman founder Prasad Thammineni explains why human adoption, not technology, is the real bottleneck, and how anyone can join the 1% by becoming an \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\"Intelligence Architect.\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\"","category":"Newsroom","segments":["Newsroom"],"tags":["Newsroom","Agent Skills","Future of Work","AI Agents","Agentic AI","Camber Creek","The Agentic Practice"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/newsroom/camber-creek-catalyst-podcast.jpg","heroImageAlt":"Prasad Thammineni, Founder \u0026 CEO of Agentman (Chain of Agents), featured on Catalyst by Camber Creek podcast.","faq":[{"question":"Will AI agents replace knowledge workers?","answer":"AI agents will automate repetitive, process-driven tasks, but they create a new role â the Intelligence Architect â for workers who codify and maintain the expertise agents rely on. The greater risk falls on those who decline to adapt rather than on the workforce as a whole. Early movers who learn to direct agents gain a durable advantage because adoption, not technology, is the current constraint."},{"question":"What is an Intelligence Architect?","answer":"An Intelligence Architect is a non-engineer who captures organizational expertise as agent-readable skill documents and maintains them as operating intelligence. The role lets people participate in the agent economy through domain knowledge and clear writing rather than coding. They decide what intelligence an agent needs and codify it, often working alongside subject-matter experts."},{"question":"Why haven't AI agents transformed most jobs yet?","answer":"The bottleneck is human adoption, not technical capability. Production-grade agent tools exist, but more than 99% of people still use AI only as a search engine, and just over 1% can build agents that run reliably in production. This knowledge gap, not a shortage of capable tools, explains the slow transformation."},{"question":"How long does it take to learn agentic technology?","answer":"A focused effort of three to six months is enough to move from beginner to the top tier of practitioners, according to Chain of Agents. The field is early, so the learning curve is shorter than the surrounding hype implies. Free training resources, including Agentman's classes, reduce the cost of starting to near zero."},{"question":"What are agent skills?","answer":"Agent skills are documents that encode a procedure, voice, or set of examples, which an agent loads at runtime to guide its behavior. They separate procedural knowledge from underlying code, so non-engineers can author and update agent behavior in a familiar editor. A self
10-learning loop lets the agent revise the skill as new patterns emerge."}]},{"slug":"governance-first-ai-audit-trails-citations-access-controls","title":"Governance-First AI: Why Audit Trails, Citations, and Access Controls Aren't Optional in Regulated Industries","date":"2026-05-20T17:07:07.985Z","readingTime":"14 min read","excerpt":"Why AI audit trails, citations, and access controls aren't optional in regulated healthcare and PE â the governance-first architecture explained, with 2025â2026 data.","category":"Agent Skills","segments":["Agent Skills"],"tags":["AI Governance","Healthcare AI","Private Equity","Compliance","Audit Trails","RegTech"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/governance-first-ai-hero.jpg","heroImageAlt":"Diagram of a governed AI agent connected to four pillars: audit trail logging, source citations, access controls, and information barriers, for regulated healthcare and finance.","faq":[{"question":"What is governance-first AI?","answer":"Governance-first AI is an architecture where audit logging, source citations, access controls, and data-handling rules are built into the AI system from the start, rather than added after deployment. In regulated industries, it means every agent action is traceable, every output is linked to a source, and every data access is scoped to the user's existing permissions."},{"question":"Do AI hallucinations really matter in healthcare and finance?","answer":"Yes. Large language models hallucinate on a majority of specific legal queries and at high rates on hard medical questions without grounding. In healthcare and finance, a confidently stated but fabricated answer can drive a clinical error, a compliance violation, or a sanctioned legal filing. Source-linked citations and retrieval grounding are the primary technical defense."},{"question":"What is a Chinese wall in the context of AI agents?","answer":"A Chinese wall, or information barrier, prevents an AI agent serving one team from accessing or surfacing another team's confidential information. In private equity, it stops an agent supporting one deal from exposing material non-public information held by a different deal team. It is enforced at the agent's data-access layer, not just in policy."},{"question":"Does Agentman use client data to train its models?","answer":"No. Agentman's governance baseline is that client data is never used for model training. This is enforced architecturally and paired with an immutable audit trail and citation engine, so clients can verify exactly how their data was used and confirm it was not absorbed into a shared model."},{"question":"How long must AI audit records be retained?","answer":"Retention depends on the regulation. HIPAA requires certain compliance documentation to be retained for at least six years, and financial services recordkeeping rules carry their own multi-year windows. Agentman matches retention periods to the governing regulation and stores records immutably so they remain audit-ready."}]},{"slug":"healthcare-network-flywheel-every-claim-makes-every-practice-smarter","title":"The Healthcare Network Flywheel: How Every Claim a Specialty Practice Processes Makes Every Other Practice Smarter","date":"2026-05-19T17:42:19.829Z","readingTime":"13 min read","excerpt":"Healthcare AI has a unique network effect that no horizontal SaaS company can replicate: when one independent specialty practice's eligibility verification agent encounters a new payer denial pattern, every practice on the platform inherits the learned pattern the same day. The Healthcare Intelligence Layer is shared infrastructure for back-office automation â and the 2025â2026 denial economics make it the most durable moat in vertical healthcare AI.","category":"Healthcare","segments":["Healthcare"],"tags":["Agentic Healthcare","Network Effects","Healthcare Intelligence Layer","Denial Management","Prior Authorization","Eligibility Verification","Revenue Cycle Management","Independent Practice","Specialty Practice","Payer Behavior","Agent Skills"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/healthcare-network-flywheel-every-claim-makes-every-practice-smarter-hero.jpg","heroImageAlt":"Flat geometric illustration of a healthcare network flywheel: six independent specialty practice nodes connected in a clockwise loop around a central intelligence hub, in Agentman's terracotta and cream palette","faq":[{"question":"What is a healthcare AI network effect?","answer":"A healthcare AI network effect is the compounding intelligence advantage that an AI platform gains as more independent specialty medical practices use it. When one practice's eligibility verification agent encounters a new payer denial pattern, the learned pattern is captured at the payer-rule level and propagated to every other practice's agent on the network. The signal compounds with practice count per specialty vertical, not raw practice count."}
10,{"question":"How is the Healthcare Intelligence Layer different from federated learning?","answer":"The Healthcare Intelligence Layer extends the federated-learning principle â bringing the model to the data without moving patient information â into back-office revenue cycle automation for independent specialty practices. Federated learning research in 2025 has focused primarily on cross-hospital AI model training for clinical outcomes. The Healthcare Intelligence Layer applies the same privacy-preserving architecture to denial patterns, payer rule changes, and prior authorization requirements."},{"question":"Does sharing denial patterns across practices violate HIPAA?","answer":"No. Agentman captures payer-rule outcomes at the CPT/ICD code pair and payer-plan identifier level, never at the level of individual patient records. No PHI crosses practice boundaries. The architecture mirrors the privacy-preserving design principles established in federated learning research and complies with HIPAA's de-identification standards."},{"question":"How much does the eligibility verification agent cost?","answer":"Agentman's eligibility verification agent costs $0.50 per check, compared to the CAQH Index industry benchmark of $7.97 per check â a 93.7% cost reduction. The economics are driven by the underlying API and inference cost, not by labor, which means the marginal cost per check declines as the network intelligence layer absorbs more payer-rule discovery work."},{"question":"Which specialties benefit most from the network flywheel today?","answer":"Wound care, vein care, diabetes \u0026 obesity, dermatology, and ophthalmology are the priority specialty verticals for Agentman's network in 2026. These specialties share three characteristics that make them ideal for the Healthcare Intelligence Layer: high prior authorization volume, distinct payer-rule terrain, and a population of independent practices large enough to generate strong per-vertical signal."}]},{"slug":"from-eligibility-to-appeal-full-revenue-cycle-automated","title":"From Eligibility to Appeal: The Full Revenue Cycle, Automated","date":"2026-05-12T03:27:17.736Z","readingTime":"12 min read","excerpt":"Independent specialty practices collect roughly 60 cents on every gross dollar billed because revenue leaks at four sequential checkpoints â eligibility, prior authorization, documentation, and denial recovery. Automating all four with agent skills closes the loop, and the 2026 industry benchmarks now make the financial case decisive.","category":"Healthcare","segments":["Healthcare","Agent Skills"],"tags":["Agentic Healthcare","Revenue Cycle Management","Eligibility Verification","Prior Authorization","Denial Management","Clinical Documentation Improvement","Independent Practice","Healthcare Economics","CMS-0057","Agent Skills"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/from-eligibility-to-appeal-hero.jpg","heroImageAlt":"Side-by-side illustration: a torn dollar bill collecting $0.60 with fragments labeled Eligibility, Prior Authorization, Documentation, and Denial Recovery, contrasted with an intact dollar bill collecting $1.00 after applying agent skills.","faq":[{"question":"What is the 60-cent dollar in healthcare?","answer":"The 60-cent dollar refers to the average gross collection rate for independent specialty practices billing commercial payers. For every dollar billed at standard charge-master rates, the practice collects approximately 60 cents after contractual adjustments, eligibility errors, prior authorization denials, documentation gaps, and unworked appeals. The remaining 40 cents includes both contractually written-off amounts and recoverable revenue that automation of the four revenue cycle checkpoints can reclaim."},{"question":"How much does manual eligibility verification cost per check in 2026?","answer":"Manual eligibility verification costs approximately $6.78 per check according to the 2025 CAQH Index, while electronic verification costs approximately $0.34 â a savings of $6.44 per transaction. For a mid-size practice, this represents thousands of dollars in daily savings on eligibility alone."},{"question":"Why do most denied claims never get resubmitted?","answer":"Approximately 35â60% of denied claims are never resubmitted because manual appeal preparation costs $25â$118 per claim and recovers only about 35â40% of the disputed amount. The ec
10onomics make appeals unattractive at the per-claim level, even though the cumulative unrecovered revenue is substantial. Agent skills change this by reducing appeal generation cost to a fraction of the manual benchmark."},{"question":"Is denial prevention different from denial management?","answer":"Yes. Denial management is the back-end process of working denials after they occur â triage, appeal letters, payer follow-up. Denial prevention is the upstream work of catching documentation, eligibility, and authorization gaps before claim submission. 86% of denials are potentially avoidable, and roughly half of avoidable denials are not recoverable once they occur â making prevention the only checkpoint where the full dollar can still be saved."},{"question":"How does CMS-0057-F change prior authorization in 2026?","answer":"CMS-0057-F took effect January 1, 2026, requiring impacted payers to respond to urgent prior authorization requests within 72 hours and standard requests within 7 calendar days, with specific denial rationale and public reporting on PA approval rates. FHIR-based Prior Authorization APIs are required by January 1, 2027. Together these provisions create the first structured, time-bound payer interaction layer that PA automation can build against."},{"question":"Does this work for practices without API access to payers?","answer":"Yes. Approximately 30â40% of payer interactions still require portal or fax workflows that lack clean API access. Agent skills handle both API-connected and portal-only payers using screen-agent technology, so the full revenue cycle is covered regardless of payer infrastructure. The 2025 CAQH Index identifies attachments as the most lagging transaction at 24% electronic adoption in medical, highlighting persistent friction for documentation exchange even as broader digitization advances."}]},{"slug":"prior-authorization-agent-skills-13-hours-to-minutes","title":"Prior Authorization Is Broken: How Agent Skills Cut 13 Hours a Week to Minutes","date":"2026-05-08T21:40:25.143Z","readingTime":"16 min read","excerpt":"Independent practice staff still lose 13 hours per physician per week to prior authorization, and 90% of practices say the burden got worse this year. Agent skills automate the full PA workflow from three trigger sources â EHR, inbox, and ambient encounter â while keeping a clinician in the loop on every submission. Here is how the mechanism works under the new CMS rules, and why \\\\\\\"clean prior auth\\\\\\\" is the part most vendors skip.","category":"Healthcare","segments":["Healthcare","Agent Skills"],"tags":["Agentic Healthcare","Prior Authorization","Agent Skills","Ambient AI","Independent Practice","Revenue Cycle Management","Practice Operations","Healthcare Automation","CMS-0057","Use Cases"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/prior-authorization-agent-skills-hero-v2.jpg","heroImageAlt":"Flat geometric illustration showing chaotic prior authorization paperwork on the left, with a powerful stream of agent skills flowing into a fully energized computer dashboard on the right â visualizing how Agentman's MedMan PA automation transforms manual document logistics into a powered, in-control workflow","faq":[{"question":"How much time does prior authorization actually consume in an independent practice in 2026?","answer":"Practices complete an average of 39 PAs per physician per week and dedicate roughly 13 hours of physician and staff time to processing them. The MGMA 2026 Regulatory Burden Report found that 90% of practices saw PA requirements increase in the past year, and 40% have hired multiple full-time administrative staff per physician to manage payer rules."},{"question":"Does an agent skill replace the clinician's judgment on prior auth?","answer":"No. An agent skill assembles the documentation and applies known payer rules, then routes the package to the clinician for approval. The clinician approves, edits, or rejects every submission. The agent handles the typing, faxing, and portal logins; the human handles the clinical judgment."},{"question":"What happens when the payer has no API?","answer":"A desktop-resident agent uses screen automation to drive the payer portal the same way a human would â logging in, completing forms, attaching documents, clicking submit. Cloud-only platforms cannot do this because their runtime sits outside the user's desktop environment."},{"question":"What changed under CMS-0057-F on January 1, 2026?","answer":"Impacted payers must now issue urgent PA decisions within 72 hours and standard decisions within 7 calendar days, must provide specific denial reasons, and began publicly reporting PA metrics in March 2026. The full FHIR-based Prior Authorization API requirement does not take effect until January 1, 2027."},{"question":"How does the agent know payer-specific approval criteria?","answer":"Payer rules are encoded as agent skills â for example, the BMI and comorbidity thresholds required by a specific plan for GLP-1 prior authorization. Each clinic that teaches the system a new rule contributes to a knowledge base every other clinic inherits."},{"question":"What is the difference between PA automation and \"clean prior auth\"?","answer":"PA automation reduces staff time on rework. Clean prior auth eliminates rework entirely by flagging missing documentation at the point of care, before the PA is submitted. Practices targeting clean submission have moved first-pass approval rates from the 60â70% industry baseline into the 90%+ range in published case studies."},{"question":"Does this work for imaging and DME, not just medications?","answer":"Yes. Medications are the most common entry point and the highest volume. Imaging â anything beyond a plain X-ray â and durable medical equipment each have dedicated skills with their own documentation patterns. Specialty procedures are added per-specialty."}]},{"slug":"prior-auth-trap-built-for-everyone-except-you","title":"Newsletter #5: The Prior Auth Trap â Built for Everyone Except You","date":"2026-05-05T21:56:58.948Z","readingTime":"16 min read","excerpt":"Physicians spend 13 hours a week on prior authorization. 67% of denials are never appealed. Every solution the industry has shipped â outsourced PA, CoverMyMeds, payer portals, AI auto-fill, the new wave of agentic PA startups â was built for the payer, not the independent practice. The first roun
10d of CMS-mandated public data confirms what staff schedules have shown for years: this is working exactly how it was designed. It just wasn't designed for us.","category":"Healthcare","segments":["Healthcare","Newsletter"],"tags":["Newsletter","The Agentic Practice","Agentic Healthcare","Prior Authorization","Independent Practice","Denial Management","CMS-0057","Payer Behavior","Revenue Cycle Management","Practice Operations","Healthcare AI"],"author":{"name":"Sachin Gangupantula","avatar":"/images/sachin-gangupantula.jpeg"},"heroImage":"/images/blog/News5_image1.png","heroImageAlt":"The Agentic Practice Issue 5 cover: The Prior Auth Trap â built for everyone except you","faq":null},{"slug":"agent-skills-legal-teams-contract-review-compliance","title":"Agent Skills for Legal Teams: Contract Review, Compliance, and the Workflows That Actually Move","date":"2026-05-04T16:00:00.000Z","readingTime":"15 min read","excerpt":"Legal teams spend 200+ hours per lawyer per year on repetitive document work. Agent skills automate contract review, compliance monitoring, legal research, and redlining with retrieval-grounded citations and human-in-the-loop sign-off â and the same architecture extends directly into engineering workflows.","category":"Legal","segments":["Legal","Agent Skills"],"tags":["Legal AI","Agent Skills","Contract Review","Compliance","Legal Automation","Use Cases","Engineering","AEO","Blog"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/agent-skills-legal-teams-hero.png","heroImageAlt":"Flat editorial illustration of legal contracts flowing through a structured grid of AI agent skill nodes into a polished contract with clause-level risk indicators","faq":[{"question":"What's the difference between an agent skill and a legal AI tool like Harvey or CoCounsel?","answer":"A legal AI tool is a vertical product â a packaged interface, a pre-trained domain, a closed feature set. An agent skill is a structured instruction set that runs inside an open agent runtime and can be cloned, customized, and audited by the firm. The two are not mutually exclusive: a firm might use a legal AI product for some workflows and skills for others, especially for the firm-specific playbook work that off-the-shelf tools do not customize deeply."},{"question":"Can agent skills replace contract attorneys or compliance officers?","answer":"No. Skills handle the rule-based, evidence-assembly portion of legal work â first-pass clause review, control evidence collection, citation lookup. The judgment work â negotiation strategy, materiality calls, attestations, advice â stays with the human. Most teams redeploy junior attorneys and analysts to higher-value work rather than reducing headcount."},{"question":"How do agent skills handle attorney-client privilege and confidentiality?","answer":"Skills run inside the firm's chosen agent runtime, with the firm's chosen data residency, encryption, and access controls. Privilege is preserved at the runtime layer â not at the skill layer â and skills are designed to operate on documents already inside privileged channels. The skill itself is just structured intelligence; confidentiality lives in the infrastructure."},{"question":"How do skills prevent the citation hallucinations that have led to sanctioned filings?","answer":"Skills prevent fabricated citations through retrieval grounding. A research skill is structurally prevented from generating a citation that doesn't exist in an authorized database â every citation in the output points to a specific case, statute, or regulation by ID. If on-point authority is not found, the skill returns a \"no authority identified\" result rather than producing a plausible-sounding fabrication."},{"question":"How long does it take to deploy a legal agent skill?","answer":"Production deployment of a legal skill typically takes 2â4 weeks, including playbook ingestion, retrieval setup, and human-in-the-loop checkpoint configuration. Most teams start with contract clause flagging â the highest-volume, highest-ROI skill â and add others over a 3â6 month rollout."},{"question":"Can our firm build our own agent skills, or do we need to clone existing ones?","answer":"Both. Skills are designed to be cloned from a registry, customized to the firm's specific playbook, and versioned independently. Most firms start by cloning a clause-flagging or research skill from myAgentSkills.ai, point it at their own playbook, and adjust the deviation thresholds. Custom-built skills follow when the workflow is firm-specific enough that no registry version applies."},{"question":"Are skills only useful for big law and large in-house teams?","answer":"No. The economics of skills favor smaller teams more, not less. A 10-attorney in-house department gets the same productivity multiplier as a 1,000-attorney firm without the infrastru
10cture overhead â because the skill is the infrastructure. Solo and small-firm practitioners are among the fastest-adopting segments in the registry. ---"}]},{"slug":"8-agent-skills-medical-billing-2026","title":"8 Agent Skills Every Medical Billing Team Needs in 2026","date":"2026-04-30T20:55:39.780Z","readingTime":"13 min read","excerpt":"The eight production-ready agent skills with the highest measurable impact on medical billing teams in 2026 â from eligibility verification to prescription refills â and the four governance guardrails that keep them safe from AI hallucinations.","category":"Healthcare","segments":["Healthcare","Agent Skills"],"tags":["Healthcare","Agent Skills","Medical Billing","AI Governance","Revenue Cycle Management","Hallucination Prevention","Education"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/8-agent-skills-medical-billing-hero.png","heroImageAlt":"Illustration of eight medical billing agent skills arranged in a grid, framed by a governance shield with audit-trail connecting lines and a human-in-the-loop reviewer.","faq":[{"question":"What's the difference between an agent skill and an AI prompt?","answer":"A prompt is a one-time instruction. A skill is a structured, versioned, governed instruction set that runs the same way every time, with full audit logs and human-in-the-loop checkpoints. Skills are how AI becomes reliable enough for regulated workflows like medical billing."},{"question":"How do agent skills handle AI hallucinations?","answer":"Skills prevent hallucinations through four structural guardrails: deterministic rule execution (the agent follows a coded procedure, not a free-form prompt), retrieval-grounded outputs (every claim cites the source record), human-in-the-loop checkpoints (every consequential action requires approval), and full data lineage (every output is logged and auditable). Unlike a chatbot, a skill cannot fabricate a payer rule, a clinical justification, or a refill protocol â if the data isn't in the record, the skill flags the gap rather than guessing."},{"question":"Do agent skills replace medical billers?","answer":"No. Skills handle the high-volume, rule-based portion of billing work â eligibility checks, claim scrubbing, posting routine payments â so billers can focus on the judgment-heavy work: complex appeals, payer relationships, and patient advocacy. Most practices redeploy billers to higher-value work rather than reducing headcount."},{"question":"How long does it take to deploy an agent skill?","answer":"Production deployment typically takes 2â4 weeks per skill, including practice-specific configuration, payer rule training, and human-in-the-loop checkpoint setup. Practices usually start with eligibility verification (the simplest, highest-ROI skill) and add others over a 3â6 month rollout."},{"question":"Are agent skills HIPAA-compliant?","answer":"Yes â when deployed inside an agent runtime that meets HIPAA requirements (BAA in place, encrypted data at rest and in transit, full audit logs, role-based access). The skill itself is just structured intelligence; HIPAA compliance lives at the runtime and infrastructure layer."},{"question":"Can our practice build our own agent skills?","answer":"Yes. Skills are designed to be cloned and customized to your practice's procedures, payer mix, and protocols. Most practices start with a registry skill (from myAgentSkills.ai), clone it, and adjust the payer rules and approval thresholds to match their workflow. ---"}]},{"slug":"eligibility-verification-foundation-practice-survival","title":"Why Eligibility Verification Is the Foundation of Practice Survival","date":"2026-04-30T19:09:59.593Z","readingTime":"10 min read","excerpt":"Manual eligibility verification consumes 90+ minutes per doctor per day and sets up a 10-week denial cascade that drains $50,000â$80,000 per provider per year. Automating this single workflow is the highest-leverage move an independent specialty practice can make â here's the math, the mechanism, and the schedule shift that turns defense into offense.","category":"Healthcare","segments":["Healthcare"],"tags":["Agentic Healthcare","Eligibility Verification","Revenue Cycle Management","Independent Practice","Denial Management","Healthcare Automation","Practice Operations","Specialty Practice"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/eligibility-verification-foundation-hero.png","heroImageAlt":"Flat-vector illustration of a calm front-desk worker at a medical practice reviewing a scheduling dashboard with completed checkmarks, while in an adjacent room a clinician has an unhurried conversation with a patient â visualizing how recovered staff hours from automated eligibility verification translate into more attentive patient care.","faq":[{"question":"What does an eligibility verification agent actually check?","answer":"An eligibility verification agent confirms active coverage, copay, deductible, plan type, network status, and prior authorization requirements for each scheduled patient. It pulls this information directly from payer systems, normalizes the response, and flags any patient whose coverage will not support the planned visit."},{"question":"How far in advance should eligibility be verified?","answer":"Eligibility should be verified five days before the appointment, not the day before. A 5-day window gives staff time to reach patients with invali
10d coverage, update insurance, or rebook the slot. Day-before verification leaves no recovery window â the slot goes empty or the visit becomes uncompensated."},{"question":"What percentage of denials trace back to eligibility issues?","answer":"Eligibility-related issues are consistently among the top three denial reasons across specialty practices, frequently driving 20-30% of total denial volume. The exact share varies by specialty and payer mix, but in every Agentman customer practice the eligibility category is large enough to be worth automating first."},{"question":"Does an eligibility agent replace front-desk staff?","answer":"No. An eligibility agent automates a single high-volume task that previously consumed 1-2 hours of staff time per provider per day. The front-desk role does not disappear â it shifts toward patient communication, scheduling optimization, and the human work that drives growth."},{"question":"How quickly does an eligibility agent pay for itself?","answer":"An eligibility agent typically pays back within the first month of deployment based on staff time saved alone. The denial-avoidance return â $50,000â$80,000 per provider per year â is realized over the following 6-10 weeks as the lagged denial cycle catches up to the new workflow."},{"question":"What payers does the Agentman eligibility agent cover?","answer":"The Agentman eligibility agent programmatically handles approximately 90% of major commercial and government payers, including the dominant carriers in most specialty markets. Payers outside this set fall back to staff review, with the agent surfacing the work and pre-filling the patient context."}]},{"slug":"doctor-didnt-ask-who-diagnosed-me","title":"The Doctor Didn't Ask Who Diagnosed Me â They Asked If the Bill Will Get Paid","date":"2026-04-28T03:31:26.197Z","readingTime":"8 min read","excerpt":"Clinical AI gets the magazine covers. Meanwhile, the administrative layer that decides whether an independent practice survives runs on fax machines, twelve browser tabs, and staff on payer hold lines. The headlines and the outcomes have drifted apart â and the joke that practice managers tell is starting to look like a strategy.","category":"Healthcare","segments":["Healthcare"],"tags":["Agentic Healthcare","Independent Practice","Revenue Cycle Management","Administrative AI","Eligibility Verification","Prior Authorization","Denial Management","Healthcare AI"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/doctor-didnt-ask-who-diagnosed-me.png","heroImageAlt":"Split flat illustration contrasting clinical AI on the left â a clean stethoscope and chest X-ray with calm geometric lines â against administrative AI on the right â a cluttered cluster of overlapping payer browser tabs, a fax machine, a phone receiver with a coiled cord, paper stacks, and a caution-amber denial-letter envelope.","faq":[{"question":"What is administrative AI in healthcare?","answer":"Administrative AI in healthcare refers to systems that automate the non-clinical workflows of running a practice: eligibility verification, prior authorization, claims preparation, denial management, payer portal monitoring, and patient communication. Unlike clinical AI, which supports diagnosis and treatment, administrative AI targets the revenue cycle and back-office operations that determine whether a practice is financially viable."},{"question":"Why do independent medical practices struggle with billing?","answer":"Independent practices struggle with billing because each payer has different rules, formularies change without notice, and there are over 72,000 ICD-10 codes with thousands more added or revised yearly. Most practices rely on the institutional memory of one or two long-tenured staff. When that knowledge walks out the door, claims get denied, payments get delayed, and the practice loses revenue it had already earned."},{"question":"How long does it take a medical practice to get paid?","answer":"The average payment cycle from patient visit to deposited payment runs roughly ten weeks for clean claims, and fourteen or more weeks when denials and appeals are involved. This is one of the longest revenue cycles of any service industry and a primary reason independent practices struggle with cash flow."},{"question":"Can AI handle medical billing better than humans?","answer":"AI handles the rules-heavy, repetitive parts of medical billing â eligibility verification, code lookup, denial pattern detection, prior auth criteria matching â more accurately and consistently than humans, because no individual can hold 72,000 codes and dozens of payer rulebooks in working memory. Humans remain essential for edge cases, patient conversations, and final claim submission."},{"question":"What is the difference between clinical AI and administrative AI?","answer":"Clinical AI supports diagnosis, imaging analysis, and treatment recommendations using structured medical data. Administrative AI automates the operational workflows that surround clinical care: insurance verification, billing, prior authorization, and revenue cycle management. Clinical AI gets more attention; administrative AI typically delivers faster ROI for independent practices. --- Special thanks to Sachin Gangupantula, VP Agentic Healthcare at Agentman and a practicing clinician at Valley Diabetes \u0026 Obesity, for the field perspective that shaped this piece."}]},{"slug":"five-minutes-to-your-first-agent-self-serve-healthcare-onboarding","t
10itle":"Why Should a Doctor Have to Book a Demo to Try Software? Inside Agentman's 5-Minute Onboarding","date":"2026-04-28T03:27:29.156Z","readingTime":"10 min read","excerpt":"Healthcare software is famously sales-led â book a demo, sit through three calls, wait weeks before anyone in the practice has touched the product. Agentman runs a different play: sign up, connect, and run your first eligibility check in five minutes, with no sales call required.","category":"Healthcare","segments":["Healthcare"],"tags":["Agentic Healthcare","Independent Practice","Self-Serve Onboarding","Eligibility Verification","Inbox Triage","Product Experience","Practice Operations","EHR Integration","Mobile-First"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/five-minutes-to-your-first-agent.png","heroImageAlt":"Isometric flat illustration of a smartphone showing an eligibility-verified result connected by a three-step floating-card flow to a stopwatch reading five minutes, in indigo and warm terracotta, representing self-serve healthcare software onboarding from sign-up to first agent in five minutes.","faq":[{"question":"How long does it take to onboard with Agentman?","answer":"Sign-up to first eligibility check takes about five minutes. The flow is self-serve: visit the website, create an account, connect to the practice management system, and run the first check. Inbox triage activates with one additional step â forwarding fax-to-email and voicemail-to-email to the unique address the platform issues at sign-up. Neither step requires a sales call, scheduled demo, or implementation engagement."},{"question":"Do I need to talk to a salesperson before trying Agentman?","answer":"No. The product is designed for self-serve activation. Practices can sign up, connect their practice management system, and run a first eligibility check without scheduling a demo or speaking to anyone at Agentman. Sales conversations happen after the practice has seen the product work, when the practice initiates a discussion about pricing, deeper integration, or expansion to additional agents."},{"question":"What is the Dropbox-style email at sign-up?","answer":"When a practice creates an account, Agentman issues a unique email address. The practice forwards their existing fax-to-email and voicemail-to-email messages to that address, which gives the inbox agent a feed to triage. The pattern mirrors how Dropbox lets users send files to a folder by email. It activates the inbox agent without requiring any new software at the practice."},{"question":"Does the 5-minute onboarding require EHR integration?","answer":"The five-minute eligibility check connects to the practice management system, not the full clinical EHR. EHR integration depth varies by vendor â some EHRs are ready on day one, others require a short implementation step. The product is built to integrate with any EHR, but practices don't need a completed EHR integration to see the eligibility agent and inbox agent work."},{"question":"Why design Agentman to work from a phone?","answer":"Small-practice owner-operators are rarely at a desk. The doctor is in exam rooms; the office manager is at the front desk or working denials. Decisions happen in available increments of time between patients. A phone-first product can be approved, escalated, or reviewed in those increments. A desktop-bound product cannot."},{"question":"How is this different from how healthcare software is usually sold?","answer":"Standard healthcare software sales follow a sales-led motion: book a demo, sit through pre-recorded walkthroughs, schedule follow-ups, wait weeks before product use begins. Agentman runs a consumer-style self-serve motion: sign up, connect, see a result in five minutes. The product earns trust by working on real data, not by being narrated in a curated demo. --- Special thanks to Prasad Thammineni (co-founder and CEO of Agentman) and Sachin Gangupantula (VP of Agentic Healthcare at Agentman and a practicing physician) for the conversations that shaped this piece. This post is adapted from the Bridging Healthcare Gaps podcast."}]},{"slug":"justify-my-care-physician-burden-agentic-systems","title":"Newsletter #4: Justify My Care â What Physicians Are Actually Saying About Burnout","date":"2026-04-22T16:34:05.210Z","readingTime":"15 min read","excerpt":"A physician I know recently said this out loud, to no one in particular, after a long day: I don't want another peer-to-peer call with insurance to justify my care. That sentence is the single most common feeling in American medicine right now â and it's not burnout. It's architecture. This is what the wheel costs, and why agentic systems are the first off-ramp in thirty years.","category":"Healthcare","segments":["Healthcare","Newsletter"],"tags":["Agentic Healthcare","Physician Burnout","Prior Authorization","Denial Management","Agentic AI","Independent Practice","Revenue Cycle Management","Ambient AI","Value-Based Care","The Agentic Practice","Newsletter"],"author":{"name":"Sachin Gangupantula","avatar":"/images/sachin-gangupantula.jpeg"},"heroImage":"/images/blog/News4_image1.png","heroImageAlt":"The Agentic Practice Issue 4 cover: the word Justify with the subtitle What physicians are actually saying about burnout","faq":[{"question":"What does \"justify my care\" mean in the context of physician burnout?","answer":"\"Justify my care\" is the phrase a physician used to describe peer-to-peer calls with insurers. It captures a third administrative layer that barely existed twenty years ago: documentation about care directed at payers, demanding that physicians prove, defend, and re-prove clinical decisions through prior authorization forms, medical necessity letters, denial appeals, and peer-to-peer calls. It is distinct from documentation for care â the chart note â which has always been part of medicine."},{"question":"Why do 67% of physicians no longer appeal prior authorization denials?","answer":"According to the AMA's 2024 National Prior Authorization Survey, 67% of physicians who don't appeal cite past experience telling them appealing wouldn't work. The non-appeal rate is the clearest signal that the denial feedback loop has broken: two-thirds of American physicians have stopped fighting back because the last hour of their day is better spent on anything else."},{"question":"How are ambient AI scribes different from agentic AI in healthcare?","answer":"Ambient AI scribes listen to the patient encounter and draft clinical notes. The physician still reviews, corrects, and signs every note, and the liability remains fully human. Agentic AI executes administrative workflows â prior authorizations, denial appeals, claims follow-up, payer communication â end-to-e
10nd, without requiring the physician to drive each step. Ambient AI compresses documentation time. Agentic AI removes the administrative layer from the physician's workflow entirely. Both can coexist: ambient captures the encounter, agentic acts on it downstream."},{"question":"Why hasn't value-based care worked for independent practices?","answer":"Value-based care arrived for large health systems that could absorb attribution models, reporting requirements, and the capital investment required to manage risk-based contracts. For independent practices, the administrative complexity of multiple VBC contracts, the cost of care management infrastructure, and the difficulty of reaching attributed-lives thresholds have kept it structurally out of reach. The physician who stayed independent specifically to practice differently is still on the same fee-for-service wheel."},{"question":"What can an independent practice do today to reduce the burden of justifying care?","answer":"Three actions are available now. First, measure the actual weekly volume of prior authorizations, denials, appeals, and peer-to-peer calls flowing through the practice, and note how much reaches the physician versus staff. Second, identify the denial categories the practice has stopped appealing â that is where recoverable revenue is sitting. Third, evaluate agentic AI systems that execute prior authorization and denial management workflows end-to-end, rather than assistive tools that only speed up individual steps. The test is simple: does the system execute the workflow, or does it just help the physician execute it faster? --- If you or a physician you know is struggling, the Physician Support Line is free, confidential, and staffed by psychiatrists: 1-888-409-0141. The Dr. Lorna Breen Heroes' Foundation also provides resources specifically for clinicians."}]},{"slug":"hospital-discharge-fax-would-your-team-find-it-today","title":"If a Hospital Discharge Fax Arrived This Morning, Would Your Team Find It Today?","date":"2026-04-17T20:23:43.923Z","readingTime":"10 min read","excerpt":"Independent practices receive 80-90 voicemails and 100-120 faxes every day. Buried somewhere in that stack is a hospital discharge notification that triggers a higher-reimbursement follow-up visit â but only if staff catch it within 48 hours. Most don't.","category":"Healthcare","segments":["Healthcare"],"tags":["Agentic Healthcare","Revenue Cycle Management","Transition Care Management","Inbox Triage","Independent Practice","Prior Authorization","Practice Operations"],"author":{"name":"Kofi Agyare-Kwabi","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/hospital-discharge-fax-transition-care-management.png","heroImageAlt":"Flat geometric illustration of a tall stack of paper documents with one critical document highlighted in indigo rising to the top, representing a priority hospital discharge fax surfacing from a busy practice inbox","faq":[{"question":"What is Transition Care Management billing?","answer":"Transition Care Management covers the services a physician provides to a patient in the 30 days following discharge from an inpatient facility. CMS requires contact with the patient within two business days of discharge and a face-to-face visit within either 7 or 14 days depending on medical complexity. TCM visits bill under CPT codes 99495 or 99496, both of which reimburse at rates significantly higher than equivalent standard office visit codes."},{"question":"How many faxes does the average single-physician practice receive daily?","answer":"According to Sachin Gangupantula, a practicing physician and VP of Agentic Healthcare at Agentman, the average single-physician independent practice receives 100 to 120 incoming faxes and 80 to 90 voicemails per day. Staff answer more than half of all incoming calls but cannot process the full fax volume in order of urgency without an automated triage system."},{"question":"Why does healthcare still rely on fax machines?","answer":"HIPAA created specific compliance provisions for fax that many electronic channels don't share, and certain payer and hospital submission workflows still mandate fax or mail for particular document types. The interoperability infrastru
10cture that would make fax obsolete â standardized APIs between hospital EHR systems and practice management systems â exists in regulatory frameworks like the CMS Interoperability Rule but is not yet consistently implemented across the payer and health system landscape."},{"question":"How does an inbox agent prioritize faxes?","answer":"The inbox agent classifies incoming messages by type and applies priority rules the practice defines. Hospital discharge notifications are flagged as critical and surface first. Prior authorization denials with active appeal windows are flagged as time-sensitive. Routine lab results and credentialing correspondence are flagged as standard. Staff review a prioritized queue rather than a chronological stack, which means the messages with the highest clinical and financial urgency reach human attention in time to act on them."},{"question":"Is this different from the blind inbox revenue problem?","answer":"The inbox challenge has two related but distinct components. The volume problem â too many messages arriving for staff to process without falling behind â drives the $50,000 annual loss to unsorted communications documented in a previous post in this series. The prioritization problem is a second layer: even with sufficient processing capacity, a system that surfaces messages chronologically rather than by urgency will consistently bury the messages that matter most. The hospital discharge fax is the clearest example of what a prioritization failure costs, in revenue and in clinical outcomes. --- Sachin Gangupantula is VP of Agentic Healthcare at Agentman and a practicing physician with nearly a decade of clinical and operational experience. Prasad Thammineni is co-founder and CEO of Agentman, former VP of Frontier AI at Salesforce."}]},{"slug":"the-baton-pass-how-ambient-ai-and-agentic-ai-complete-each-other","title":"Ambient AI vs Agentic AI in Healthcare: The Baton Pass That Closes the Prior Authorization Loop","date":"2026-04-17T01:03:24.794Z","readingTime":"13 min read","excerpt":"Ambient AI captures the clinical encounter; agentic AI acts on it. The integration between them â the baton pass â is where 13 hours of staff time per physician per week is lost, and where the next wave of healthcare automation is being built.","category":"Healthcare","segments":["Healthcare"],"tags":["Agentic Healthcare","Prior Authorization","Agentic AI","Ambient AI","Revenue Cycle Management","Healthcare Integration","Independent Practice"],"author":{"name":"Kofi Agyare-Kwabi","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/physician-exits-room-prior-auth-already-running.png","heroImageAlt":"Flat geometric illustration of a relay race baton pass representing the handoff from ambient AI clinical documentation to agentic AI administrative execution in healthcare","faq":[{"question":"What is the difference between ambient AI and agentic AI in healthcare?","answer":"Ambient AI converts clinical encounters into structured documentation â the encounter note, problem list, and coding hints. Agentic AI receives signals from that documentation and executes downstream administrative workflows like prior authorizations, eligibility checks, denial appeals, and claim submissions. Ambient captures the encounter; agentic acts on it. They are complementary layers in the healthcare automation stack, not competing products."},{"question":"Why does prior authorization still require manual work if the encounter note already contains the clinical information?","answer":"Because the integration between the documentation layer and the execution layer has not been standardized. Encounter notes are generated as finished artifacts for human readers, not as event streams for agents. Staff manually scan the note, identify actionable clinical events, look up payer coverage rules, and re-enter information into portal forms. The baton pass â converting a signed clinical event into an automated agent trigger â is the missing piece."},{"question":"Are prior authorizations submitted automatically once the agent assembles them?","answer":"No. Agents prepare the submission and queue it for human review. A billing or clinical staff member verifies accuracy and approves before anything is filed with a payer. The practice retains full clinical and compliance accountability. What changes is not the approval authority but the assembly work â staff stop doing portal data entry and start doing final review at much higher volume."},{"question":"How much time does a typical specialty practice spend on prior authorization today?","answer":"According to the 2024 AMA Prior Authorization Physician Survey of 1,000 practicing physicians, each physician's practice spends 13 hours per week on prior authorization and completes 39 PA requests per physician per week. Forty percent of physicians have staff working exclusively on prior authorizations. For a five-physician specialty practice, that is roughly 65 hours of staff time per week directed at a single administrative workflow."},{"question":"Does agentic AI replace the ambient AI platform?","answer":"No. Agentic AI depends on ambient AI as its upstream input. The encounter note produced by the ambient platform is the trigger that activates the agentic workflows. Practices that have already deployed ambient AI have completed the hardest part of the integration â the structured clinical documentation â and agentic AI extends the value of that investment into the back-office operations where administrative time is lost."},{"question":"Which payer-specific skills matter most for independent specialty practices?","answer":"The highest-volume skills depend on specialty, but common anchors include GLP-1 prior authorization (endocrinology, primary care), advanced imaging pre-authorization (orthopedics, oncology, neurology), biologics authorization (rheumatology, dermatology, GI), DME coverage verification (pulmonology, sleep, orthopedics), and specialty pharmacy step-therapy documentation. Skills are payer-specific because the rules and formats differ across Aetna, UnitedHealthcare, Blue Shield plans, and Medicare Advantage â and the same drug or procedure can require different evidence for different payers."}]},{"slug":"sixty-cent-dollar-independent-practice-economics","title":"The 60-Cent Dollar: The Brutal Economics Behind Independent Practice Survival","date":"2026-04-13T08:00:00.000Z","readingTime":"7 min read","excerpt":"For every dollar billed to insurance, independent practices collect only 60 cents â and they may wait 10 weeks to collect it. Understanding the two-bucket problem is the first step to fixing the math.","category":"Healthcare","segments":["Healthcare"],"tags":["Agentic Healthcare","Healthcare Economics","Independent Practice","Revenue Cycle Management","Eligibility Verification","Prior Authorization","Practice Operations"],"author":{"name":"Kofi Agyare-Kwabi","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/sixty-cent-dollar-independent-practice-economics.png","heroImageAlt":"Flat design illustration showing a dollar bill split into 60 cents and 40 cents, representing the revenue gap in independent medical practice billing","faq":[{"question":"Why do independent practices only collect 60 cents on the dollar?","answer":"The gap reflects contractual adjustments between b
10illed charges and payer-contracted rates, claim denials, underpayments, and patient balances that age beyond collectibility. The most preventable portion, driven by eligibility gaps, prior authorization failures, and coding errors, is where AI agents have the most direct impact."},{"question":"What is the two-bucket problem?","answer":"The two-bucket problem refers to the two streams of accounts receivable in a medical practice: money patients owe and money payers owe, each aging differently and requiring different management. Active management of both buckets before they age is the core function of revenue cycle management."},{"question":"Can AI agents reduce administrative costs for a small practice?","answer":"Reallocation is more accurate than elimination here. A staff member previously spending four hours a day on payer portal logins can redirect that time to patient care and complex cases requiring judgment, with the financial gain coming from eliminating the $40,000 to $50,000 annual cost of a dedicated eligibility resource, reducing denial discovery work, and stopping the write-offs from unworked items that age out silently."},{"question":"How do I know if my practice is losing revenue to administrative gaps?","answer":"Two signals matter most: first-pass denial rate and A/R aging. Industry averages run 5 to 10% for in-network claims and up to 22% for non-contracted payers, so if your practice is above those ranges, a significant portion of the gap is administrative rather than clinical. If more than 20% of insurance A/R is over 90 days, you have a compounding collection problem, and an eligibility or inbox audit typically surfaces the specific breakpoints within days. --- Sachin Gangupantula is VP of Agentic Healthcare at Agentman and a practicing physician with nearly a decade of clinical and operational experience. Prasad Thammineni is co-founder and CEO of Agentman, former VP of Frontier AI at Salesforce. This post is adapted from the Bridging Healthcare Gaps podcast. Listen to the full episode â"}]},{"slug":"what-is-an-intelligence-architect","title":"What Is an Intelligence Architect â and Why Is It the Most Important Role in AI?","date":"2026-04-12T10:00:00.000Z","readingTime":"10 min read","excerpt":"The Intelligence Architect codifies domain expertise into structured AI skills that compound with every use. This emerging role is reshaping knowledge work â here's the framework and how to become one.","category":"Platform","segments":["Learn","Platform"],"tags":["Intelligence Architect","Future of Work","AI Skills","Agentic AI","Knowledge Work","Prompt vs Skill","Agent Skills"],"author":{"name":"Agentman Academy","avatar":"/images/agentman-logo.png"},"heroImage":"/images/blog/agentman-academy-hero.png","heroImageAlt":"The Intelligence Architect â codifying domain expertise into structured AI skills that compound with every use","faq":null},{"slug":"blind-inbox-bleeding-practice-dry","title":"Newsletter #3: The Blind Inbox That's Bleeding Your Practice Dry","date":"2026-04-07T08:00:00.000Z","readingTime":"11 min read","excerpt":"Independent medical practices lose over $50,000 annually to unsorted faxes, voicemails, and unworked messages. An AI inbox triage agent can cut processing time from 120 minutes to 15 minutes per day and recover revenue hiding in the blind inbox.","category":"Healthcare","segments":["Healthcare","Newsletter"],"tags":["Agentic Healthcare","Inbox Triage","Independent Practice","Revenue Cycle Management","Fax Automation","Prior Authorization","Practice Operations","The Agentic Practice","Newsletter"],"author":{"name":"Sachin Gangupantula","avatar":"/images/sachin-gangupantula.jpeg"},"heroImage":"/images/blog/blind-inbox-bleeding-practice-dry.jpg","heroImageAlt":"AI inbox triage agent organizing faxes, voicemails, and messages for an independent medical practice","faq":[{"question":"How many faxes does a typical independent practice receive per day?","answer":"Most independent practices receive 30 to 50 faxes per day. Specialty practices handling prior authorizations, referrals, and pharmacy communications tend toward the higher end. Valley Diabetes \u0026 Obesity processes approximately 47 faxes daily alongside voicemails and emails. The Documo 2025 survey found more than a third of all inbound healthcare documents are still faxes."},{"question":"What is the biggest revenue risk from an unsorted medical practice inbox?","answer":"Prior authorization denials with expiring appeal windows represent the highest immediate revenue risk. When a PA denial sits unread, the appeal window runs, the patient's treatment is interrupted, and the practice loses both the reimbursement and the follow-up visit revenue. Expired referral authorizations and missed discharge follow-ups are the second and third largest sources of silent revenue loss."},{"question":"Can an AI inbox triage agent integrate with my EHR?","answer":"Current AI inbox triage agents operate as a classification and routing layer â they read inbound faxes, voicemails, and emails, then prioritize and assign actions. Direct EHR read-write integration is the next frontier in agentic healthcare. The immediate value comes from ensuring no item goes unworked, unclassified, or unrouted, regardless of EHR connectivity."},{"question":"How long does it take to deploy an AI inbox triage agent?","answer":"Deployment timelines vary by practice size and inbound volume, but the core setup â connecting fax, email, and voicemail channels to the AI agent â can be completed in days, not months. The agent begins learning document patterns immediately and reaches high classification accuracy within the first week of processing real practice communications."},{"question":"Is healthcare fax going away?","answer":"No. Despite CMS's 2018 goal to eliminate fax by 2020, MGMA reports 89% of healthcare organizations still use fax as of recent polling. The industry consensus has shifted from eliminating fax to making fax intelligent through AI-powered classification and routing. MGMA found 68% of practices added or expanded AI use in 2025, with the strongest adoption in inbound communications workflows."}]},{"slug":"why-healthcare-ai-that-only-works-with-apis-is-incomplete","title":"Why Healthcare AI That Only Works With APIs Is Missing 40% of the Problem","date":"2026-04-06T12:00:00.000Z","readingTime":"9 min read","excerpt":"Every healthcare AI startup building prior auth automation assumes payers have APIs. They don't â an estimated 30â40% of payer portal interactions lack clean API access. Cloud-only platforms can automate 60â70% of the revenue cycle. Medman's screen agent te
10chnology closes the rest.","category":"Healthcare","segments":["Healthcare"],"tags":["Agentic Healthcare","Prior Authorization","Revenue Cycle Management","Payer Portal Automation","AI Agents","Technical Architecture","Independent Practice"],"author":{"name":"Kofi Agyare-Kwabi","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/healthcare-ai-api-gap-40-percent.png","heroImageAlt":"Flat illustration showing two paths from a medical practice â one through clean API connections and another through messy browser portals â with a desktop application bridging both","faq":[{"question":"Does Agentman work with payers that don't have APIs?","answer":"Yes. Medman's screen agent navigates payer portals directly â any portal a human staff member can access, Agentman can automate. This covers the estimated 30â40% of payer interactions where clean APIs don't exist, including regional payers, state Medicaid systems, IPA portals, and specialty payer networks."},{"question":"What is Medman?","answer":"Medman is Agentman's desktop application â the command center for small practice back-office operations. It combines API-based automation for payers with clean integrations and screen agent technology for portal-only payers, providing 100% payer coverage from a single interface."},{"question":"How is a screen agent different from RPA?","answer":"Traditional RPA records and replays click sequences, which break when portal interfaces change. Screen agents operate at the desktop level using visual recognition and intelligent navigation â they adapt to UI changes, handle CAPTCHA challenges, and manage authentication flows that RPA tools cannot."},{"question":"Can cloud-only competitors add this capability?","answer":"Not without a fundamental architecture rebuild. Screen agent technology requires a desktop application that controls the full screen â this is an architectural choice, not a feature that can be added to a cloud-only platform through an API update or plugin."},{"question":"Is the screen agent HIPAA-compliant?","answer":"Every action the screen agent takes is logged with the same data lineage and audit trail that Agentman applies to all transactions. Access is scoped per user and role, data is encrypted in transit and at rest, and the platform is designed for HIPAA compliance from the architecture level."}]},{"slug":"5-billion-problem-small-practices-left-behind-healthcare-ai","title":"The $5B Problem No One's Solving: Why Small Practices Are Left Behind by Healthcare AI","date":"2026-04-03T12:00:00.000Z","readingTime":"13 min read","excerpt":"The healthcare AI market has poured billions into large health systems while leaving ~190,000 small practices to drown in the same administrative complexity â prior authorizations, claim denials, payer variability â without the resources to fight back. The $5B RCM opportunity for independent clinics remains largely unaddressed.","category":"Healthcare","segments":["Healthcare","Agent Skills"],"tags":["Agentic Healthcare","Small Practice","Independent Practice","Revenue Cycle Management","AI Agents","Agent Skills","Market Analysis"],"author":{"name":"Kofi Agyare-Kwabi","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/small-practices-left-behind-healthcare-ai.png","heroImageAlt":"Flat illustration showing a large hospital surrounded by AI technology while a small family medicine clinic is left in shadow, representing the gap in healthcare AI adoption","faq":[{"question":"How much does it cost a small practice to use AI for RCM?","answer":"Agentman's pricing starts at $0.50 per eligibility check on a pay-as-you-go basis, with no contract required. Subscription plans begin at $150 per physician per month for unlimited checks. A typical two-physician practice saves $5,000â$6,000 per month compared to manual verification costs, making the ROI clear within the first week of use."},{"question":"Do small practices need an IT department to deploy AI agents?","answer":"No. Agentman's agents work without EHR integration through manual upload on day one. Practices on AdvancedMD or DrChrono can connect with one click. No IT staff, no multi-month implementation, no professional services fees required."}
10,{"question":"How is this different from outsourcing billing to an offshore company?","answer":"Outsourced billing companies provide manual labor at lower cost â human staff processing claims and following up on denials. AI agents execute the same workflows in seconds with full auditability, 24/7 availability, and zero variance. Every decision is logged, traceable, and improvable. Offshore teams can't provide data lineage for every claim action."},{"question":"Are AI agents safe for handling patient health information?","answer":"Agentman owns the full technology stack â no third-party AI wrappers. Every agent action is logged, access is scoped per user and role, and data is encrypted in transit and at rest. The platform is designed for HIPAA compliance from the architecture level, with audit trails that support payer and regulatory reviews."},{"question":"Which specialties benefit most from healthcare AI agents?","answer":"Specialties with high prior authorization volume â orthopedics, cardiology, endocrinology, pain management, and vascular â see the fastest ROI because the administrative burden per patient is highest. Primary care and family medicine practices benefit from sheer volume â 25+ patients per day per physician generates hundreds of eligibility checks weekly that agents can automate."}]},{"slug":"43-billion-eligibility-verification-tax-independent-practice-automation","title":"Newsletter #2: 12 Payer Portals, Zero Logins: The $43 Billion Administrative Tax Nobody Talks About","date":"2026-03-30T12:00:00.000Z","readingTime":"14 min read","excerpt":"The U.S. healthcare industry spends $43 billion annually on eligibility and benefit verification â the most expensive administrative workflow in medical practice operations. Most of this work is still performed manually, with front-desk staff logging into 12 separate payer portals per session. A single-agent automation approach can eliminate portal logins entirely while improving verification accuracy.","category":"Healthcare","segments":["Healthcare","Newsletter"],"tags":["Agentic Healthcare","Revenue Cycle Management","Eligibility Verification","Payer Portal Automation","Independent Practice","Administrative Costs","The Agentic Practice","Newsletter"],"author":{"name":"Sachin Gangupantula","avatar":"/images/sachin-gangupantula.jpeg"},"heroImage":"/images/blog/43-billion-admin-tax-hero.jpg","heroImageAlt":"Infographic showing six key statistics about the $43 billion administrative tax in healthcare eligibility verification â industry scale, cost gap, and denial impact","faq":[{"question":"How many payer portals does a typical independent practice need to access for eligibility verification?","answer":"A typical independent practice accesses 12 or more payer portals per verification session, spanning Medicare, Medicaid, commercial insurance plans, and IPA networks. Each portal requires separate login credentials, has a different user interface, and imposes its own quirks â from session timeouts to mandatory password resets. This portal fragmentation is one of the primary drivers of the $43 billion annual industry spend on eligibility verification. MGMA's own survey data confirms this reality â most practices already know they're juggling too many payer portals, with portal overload compounding the eligibility and prior authorization burden that drains staff time and revenue."},{"question":"What is the cost difference between manual and automated eligibility verification?","answer":"Manual eligibility verification costs $6.78 per transaction compared to $0.34 for a fully electronic verification â a 20x gap. For a practice performing 50 verifications per day, manual processing costs approximately $339 daily ($83,720 annually) versus $17 daily ($4,420 annually) for automated checks. These figures represent direct transaction costs and do not include downstream costs from claim denials, rework, or lost patients."},{"question":"Why can't EHR systems handle eligibility verification adequately?","answer":"EHR built-in eligibility checks confirm only whether coverage is active or inactive. They typically cannot extract copay amounts, deductible status, out-of-pocket maximums, in-network vs. out-of-network status for specific plans, or PCP assignment verification. They also struggle with Medicaid eligibility resolution, local IPA plans, and mismatched patient data. These gaps mean staff must still access payer portals directly for the information needed to prevent claim denials."},{"question":"How many times should eligibility be verified per patient encounter?","answer":"Industry best practice requires a minimum of three verification touchpoints per patient encounter: at scheduling (confirming coverage and network status), 48 to 72 hours before the appointment (catching coverage changes), and at check-in (final confirmation and copay collection). For chronic care patients seen monthly or biweekly, verification also triggers with every referral and prior authorization. A practice seeing 30 to 40 patients daily generates 90 to 120 verification events per day."},{"question":"What happens to the verification process when the primary staff member is unavailable?","answer":"When the staff member responsible for eligibility verification is absent â due to illness, vacation, or turnover â institutional knowledge about portal-specific workarounds, payer-specific requirements, and process exceptions leaves with them. The upcoming schedule begins to deteriorate because replacement staff may
10not know which portals require phone calls, which Medicaid plans need manual lookup, or which IPA networks have non-standard verification procedures. This single-point-of-failure risk is one of the strongest arguments for automation."}]},{"slug":"healthcare-ai-adoption-2026-survey-data","title":"75% of Health Systems Have Adopted AI â Where Is the Whitespace Left?","date":"2026-03-27T20:00:00.000Z","readingTime":"15 min read","excerpt":"New survey data from 120 health systems reveals that three-quarters have deployed at least one AI solution, with multi-solution adoption growing 67% year-over-year. But the real story is in the categories where demand is massive and implementation is still low â exactly where agentic healthcare platforms are positioned to win.","category":"Healthcare","segments":["Healthcare"],"tags":["Agentic Healthcare","AI Adoption","Revenue Cycle Management","Eligibility Verification","Denial Management","Prior Authorization","Health Systems","Market Data"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/healthcare-ai-adoption-2026.png","heroImageAlt":"Healthcare AI adoption data visualization showing 75% of health systems have adopted at least one AI solution in 2026","faq":[{"question":"How many health systems have adopted AI in 2026?","answer":"According to the Eliciting Insights February 2026 survey of 120 health systems, 75% have deployed at least one AI solution. Of those, 59% are running three or more AI solutions simultaneously, representing 67% year-over-year growth in multi-solution adoption."},{"question":"Which healthcare AI category has the highest ROI?","answer":"AI-based clinical documentation improvement (CDI) shows the highest ROI among implemented solutions, with 71% of implementers reporting 2x or higher returns. AI denial prediction is second at 70%, with 43% of denial prediction implementers reporting 3x or greater returns."},{"question":"What healthcare AI solutions are health systems considering in 2026?","answer":"The highest \"actively considering\" rates are in revenue cycle management: AI agents for eligibility verification at 45%, AI denial prediction at 44%, and AI clinical decision support at 43%. These categories show the largest gap between demand (consideration) and deployment (implementation), signaling imminent buying decisions."},{"question":"How does health system size affect AI adoption?","answer":"Large health systems (900+ beds) show 91% adoption of ambient listening and 55% adoption of eligibility AI agents. Sub-100 bed systems show 60% ambient listening adoption but near-zero adoption of RCM-specific AI. Receptivity scores range from 45 (sub-100 beds) to 65 (900+ beds), indicating willingness but not access."},{"question":"What percentage of health systems use Epic as their EMR?","answer":"Epic holds 48% of the health system EMR market, followed by Oracle/Cerner at 16%, Other systems at 17%, Meditech at 8%, CPSI at 7%, and AthenaHealth at 4%. AI adoption rates are consistently higher on Epic systems across all categories."}]},{"slug":"newsletter-1-independent-practice-agentic-ai-fight-back","title":"Newsletter #1: The Healthcare System Wasn't Built for You. Now You Can Fight Back","date":"2026-03-23T12:00:00.000Z","readingTime":"11 min read","excerpt":"Independent medical practices face a five-front squeeze â collapsing reimbursement, crushing administrative burden, AI-powered claim denials, staff burnout, and technology that solves one problem while worsening four others. Agentic AI gives independent practices the first tool capable of fighting back on every front simultaneously.","category":"Healthcare","segments":["Healthcare","Newsletter"],"tags":["Agentic Healthcare","Revenue Cycle Management","Agentic AI","Independent Practice","Claim Denials","Prior Authorization","The Agentic Practice","Newsletter"],"author":{"name":"Sachin Gangupantula","avatar":"/images/sachin-gangupantula.jpeg"},"heroImage":"/images/blog/newsletter-1-independent-practice-agentic-ai-hero.jpg","heroImageAlt":"Infographic illustrating the five-front squeeze on independent medical practices â
10collapsing reimbursement, administrative burden, AI-powered denials, staff burnout, and fragmented technology","faq":[{"question":"What is agentic AI in healthcare?","answer":"Agentic AI in healthcare refers to AI systems that plan multi-step workflows, take action across platforms like EHRs, clearinghouses, and payer portals, and coordinate complex operational processes with human oversight. Unlike chatbots or documentation assistants that handle single tasks, agentic AI manages entire operational workflows end to end â from eligibility verification through denial management."},{"question":"How is agentic AI different from AI medical scribes?","answer":"AI medical scribes automate one task: clinical documentation. Agentic AI coordinates across the full revenue cycle â eligibility verification, prior authorization, claim submission, denial discovery, and denial management. Scribes reduce documentation time but don't address the 65% of denied claims that never get resubmitted or the 13 hours per week staff spend on prior authorization calls."},{"question":"Can small practices afford agentic AI?","answer":"Agentic AI platforms built specifically for independent practices price for 1â50 provider operations, not enterprise budgets. The ROI calculation is straightforward: if a practice loses revenue to avoidable denials (86% of all denials are avoidable), recovering even a fraction of currently abandoned claims â the 65% never resubmitted â generates immediate, measurable return."},{"question":"What should independent practices automate first?","answer":"Eligibility verification is the highest-impact starting point. Eligibility errors cascade through the entire revenue cycle, causing claim denials, delayed reimbursement, and wasted staff time on preventable corrections. Automating eligibility at the point of scheduling eliminates a class of downstream problems before they ever occur. Prior authorization and denial management follow as high-priority next steps."},{"question":"Are payers using AI against independent practices?","answer":"71% of insurers now use AI for utilization management. One major insurer's algorithm processed 300,000 claim denials in two months, with reviewers spending an average of 1.2 seconds per case. Denial rates rose to 11.8% in 2024. Payers are deploying AI at scale to deny claims. Independent practices using manual processes â fax machines, phone trees, and understaffed billing departments â are structurally outmatched. ---"}]},{"slug":"agent-skills-flywheel-every-clinic-smarter","title":"Agent Skills Are the Flywheel That Makes Every Clinic Smarter","date":"2026-03-18T00:00:00.000Z","readingTime":"9 min read","excerpt":"Agent skills create a compounding network effect in healthcare AI. When one clinic teaches an agent how to handle Blue Shield GLP-1 prior authorizations, every similar clinic on the platform inherits that knowledge. The more clinics that deploy, the smarter every agent gets.","category":"Healthcare","segments":["Healthcare","Agent Skills"],"tags":["Agentic Healthcare","Agent Skills","Prior Authorization","Network Effects","Workflow Automation","AI Agents","Independent Practice"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/agent-skills-flywheel-every-clinic-smarter-hero.png","heroImageAlt":"Flat design illustration of a healthcare AI flywheel with interconnected clinic nodes sharing knowledge across a network","faq":[{"question":"What is an agent skill in healthcare AI?","answer":"An agent skill is a structured set of instructions written in plain English that tells an AI agent how to perform a specific clinical or administrative process. Skills are versioned, editable, and improve through a reinforcement learning feedback loop. They capture tribal knowledge â like payer-specific prior authorization requirements â and make it executable and transferable across practices."},{"question":"How do agent skills differ from traditional AI prompts?","answer":"Prompts are one-time instructions that produce variable outputs and do not learn from execution. Agent skills are persistent, versioned assets that accumulate institutional knowledge over time. Skills follow specific templates, maintain payer and regional metadata, and improve through production feedback loops. The distinction matters most at scale â prompts drift, skills compound."},{"question":"Can a small clinic benefit from the skills flywheel?","answer":"Small clinics benefit disproportionately. A five-provider practice can license production-tested skills from the marketplace instead of spending months building custom workflows. The skills they adopt have already been refined by larger practices processing higher volumes â meaning small clinics get enterprise-grade intelligence at a fraction of the cost."},{"question":"How does the skills marketplace work?","answer":"Clinics that build and refine high-quality skills can license them to other practices for a monthly fee. The originating practice monetizes its operational expertise while receiving practices skip months of trial and error. Skills in the marketplace carry production success metrics, so buyers can evaluate quality before licensing."},{"question":"What happens when payer rules change?","answer":"Skills update as rules change. When a payer modifies its formulary or prior authorization requirements, the feedback loop detects failures in execution, triggers a skill revision, and propagates the update across every practice using that skill. Static knowledge bases require manual updates and often go stale within weeks."}]},{"slug":"valley-diabetes-case-study-ai-agent-skills-results","title":"How a Small Specialty Practice Achieved 90% Automation and $149K Annual Savings with AI Agent Skills","date":"2026-03-17T00:00:00.000Z","readingTime":"9 min read","excerpt":"Valley Diabetes \u0026 Obesity deployed Agentman's healthcare agent skills and achieved 90% automation of eligibility and inbox workflows, reduced denials by 65%, and projected $107Kâ$149K in annual savings per physician â all within 12 weeks.","category":"Healthcare","segments":["Healthcare","Agent Skills"],"tags":["Agentic Healthcare","Case Study","Eligibility Verification","Revenue Cycle Management","Independent Practice","Agent Skills","Prior Authorization"],"author":{"name":"Sachin Gangupantula","avatar":"/images/sachin-gangupantula.jpeg"},"heroImage":"/images/blog/hero image.png","heroImageAlt":"Valley Diabetes \u0026 Obesity case study showing 90% automation and $149K annual savings with Agentman AI agent skills","faq":[{"question":"How long does it take to deploy the first Agentman agent skill?","answer":"That depends on what \"deployed\" means. A practice can sign up, connect its practice management system, and run its first live eligibility check in about five minutes â no s
10ales call, no implementation project. That first check is real, not a demo. Full production coverage across a practice's entire payer mix is a longer process. VDO began its engagement in July 2025, deployed its first agent skill â eligibility verification â in November 2025, and was handling production workflows within the first week of go-live. Most practices can expect the first skill to reach full operational coverage across their complete payer mix within 30â60 days of kickoff, depending on payer mix complexity. The 30â60 day window is about breadth (every payer, every edge case, fully trusted by staff), not about time-to-first-result."},{"question":"Does Agentman replace billing staff?","answer":"No. Agentman's agent skills are designed as human-in-the-loop systems. Staff review exceptions and exercise judgment on complex cases; the agent handles high-volume routine tasks. VDO's staff now spend their time on patient engagement and higher-complexity billing â not repetitive portal logins."},{"question":"What size practice benefits most from healthcare AI agent skills?","answer":"Small to mid-size specialty practices with 1â10 physicians see the fastest ROI because manual back-office burden is highest relative to staffing capacity. VDO, a specialty practice in Central California, achieved $107Kâ$149K in projected annual savings per physician within 12 weeks."},{"question":"How does the phased deployment model work?","answer":"Each agent skill is deployed independently, starting with the highest-friction workflow. The skill must prove measurable results before the next one is introduced. VDO followed this sequence: eligibility verification first, then inbox triage, with prior authorization next. Each phase builds on the trust and capacity gains established by the previous one."},{"question":"What payer systems does the eligibility agent support?","answer":"The eligibility agent runs automated checks across all active payer contracts â including Medicare, Medicaid, commercial insurers, and IPA portals. VDO's environment included 10â15 payer contracts across 12+ portal systems, all handled by a single agent skill. ---"}]},{"slug":"ic-memo-that-wrote-itself","title":"The IC Memo That Wrote Itself: How One Firm Saved 10+ Hours on a Single Deal","date":"2026-03-13T21:21:53.125Z","readingTime":"7 min read","excerpt":"IC memo preparation consumes partner weekends. One PE firm used Agentman's IC Memo Generator skill chain to cut memo prep from 15 hours to 2 â with better citations and firm-specific formatting.","category":"PE","segments":["PE","Agent Skills"],"tags":["Private Equity","Agent Skills","Use Cases","Blog"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/PE_blog.png","heroImageAlt":"Flat illustration of scattered deal documents flowing through a structured AI skill chain into a polished IC memo","faq":null},{"slug":"how-agent-skills-are-changing-deal-diligence","title":"How Agent Skills Are Changing Deal Diligence from Days to Minutes","date":"2026-03-13T20:48:26.505Z","readingTime":"6 min read","excerpt":"A typical CIM analysis takes 8â12 analyst hours. An agent skill does it in minutes â with structured extraction, risk flagging, thesis matching, and cited IC memos.","category":"PE","segments":["PE","Agent Skills"],"tags":["Private Equity","Agent Skills","Use Cases","Deal Diligence","CIM Analysis"],"author":{"name":"Debby Wang","avatar":"/images/debby-wang.jpeg"},"heroImage":"/images/blog/cim-blog.png","heroImageAlt":"Flat design illustration showing a CIM document being analyzed by AI agent skills, with data flowing into a structured IC memo","faq":null},{"slug":"ai-skills-for-healthcare","title":"AI Skills for Healthcare: From Prior Auth to Clinical Documentation","date":"2026-03-12T10:00:00.000Z","readingTime":"6 min read","excerpt":"Healthcare teams waste hours on prior authorizations, clinical notes, and compliance checks. AI skills encode your protocols, payer rules, and documentation standards so every task meets your organization's quality bar.","category":"Healthcare","segments":["Learn","Healthcare"],"tags":["Healthcare","Clinical Documentation","Prior Authorization","HIPAA","Industry Guide","Agent Skills","AI Skills"],"author":{"name":"Agentman Academy","avatar":"/images/agentman-logo.png"},"heroImage":"/images/blog/agentman-academy-hero.png","heroImageAlt":"AI Skills for Healthcare â prior authorization, clinical documentation, and compliance encoded as reusable skills","faq":null},{"slug":"ai-skills-for-sales","title":"AI Skills for Sales: Qualification, Outreach \u0026 Deal Intelligence","date":"2026-03-11T10:00:00.000Z","readingTime":"7 min read","excerpt":"Sales teams lose deals because qualification is inconsistent, outreach is generic, and tribal knowledge disappears with turnover. AI skills encode your top rep's playbook so every rep sells like your best.","category":"Agent Skills","segments":["Learn","Agent Skills"],"tags":["Sales","Lead Qualification","Sales Outreach","Deal Intelligence","Industry Guide","Agent Skills","AI Skills"],"author":{"name":"Agentman Academy","avatar":"/images/agentman-logo.png"},"heroImage":"/images/blog/agentman-academy-hero.png","heroImageAlt":"AI Skills for Sales â qualification frameworks, outreach sequences, and deal intelligence encoded as reusable skills","faq":null},{"slug":"ai-skills-for-marketing","title":"AI Skills for Marketing: Brand Voice, Campaign Briefs \u0026 Content at Scale","date":"2026-03-11T09:00:00.000Z","readingTime":"8 min read","excerpt":"Marketing teams struggle with brand consistency across channels and team members. AI skills encode your brand voice, campaign frameworks, and content standards so every piece of content sounds like it came from your best writer.","category":"Agent Skills","segments":["Learn","Agent Skills"],"tags":["Marketing","Brand Voice","Content Marketing","Campaign Strategy","Industry Guide","Agent Skills","AI Skills"],"author":{"name":"Agentman Academy","avatar":"/images/agentman-logo.png"},"heroImage":"/images/blog/agentman-academy-hero.png","heroImageAlt":"AI Skills for Marketing â brand voice, campaign briefs, and content creation encoded as reusable skills","faq":null},{"slug":"what-are-ai-skills-complete-guide","title":"What Are AI Skills? The Complete Guide for 2026","date":"2026-03-10T10:00:00.000Z","readingTime":"2 min read","excerpt":"Everything you need to know about AI skills â what they are, how they work, and why they're replacing ad-hoc prompting across enterprises.","category":"Agent Skills","segments":["Learn","Agent Skills"],"tags":["AI Skills","Getting Started","MCP","Enterprise AI","Agent Skills","Education"],"author":{"name":"Agentman Academy","avatar":"/images/agentman-logo.png"},"heroImage":"/images/blog/agentman-academy-hero.png","heroImageAlt":"Agentman Academy â what are AI skills, the complete guide","faq":null},{"slug":"small-practice-primer-for-vcs","title":"A Small Practice Primer for VCs â Why the Back-Office Is the Opportunity You're Missing","date":"2026-03-07T18:59:37.734Z","readingTime":"12 min read","excerpt":"Nine months ago, I didn't understand how a small medical practice actually works. Then one analogy changed everything. This is the primer I wish someone had given me â and the one I now give to every VC who asks why we're building Agentman.","category":"Healthcare","segments":["Healthcare"],"tags":["Agentic Healthcare","Small Practice","Independent Practice","Revenue Cycle Management","Venture Capital","Business Model","Healthcare Economics"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/practice-is-a-hotel.png","heroImageAlt":"Side-by-side illustration comparing a hotel and a small medical practice, showing parallel operational workflows","faq":null},{"slug":"a-prompt-is-a-question-a-skill-is-a-procedure","title":"A Prompt Is a Question. A Skill Is a Procedure.","date":"2026-03-05T21:03:50.580Z","readingTime":"7 min read","excerpt":"Your team's best prompt is decomposing in a Slack thread. Skills encode domain knowledge, procedures, and validation into assets your organization actually owns.","category":"Platform","segments":["Learn","Platform","Agent Skills"],"tags":["AI Skills","Agent Skills","Claude Skills","Enterprise AI","Prompt Engineering","Structured AI","myAgentSkills","MCP","Claude Code"],"author":{"name":"Agentman Academy","avatar":"/images/agentman-logo.png"},"heroImage":"/images/blog/agentman-academy-hero.png","heroImageAlt":"Agentman Academy â a prompt is a question, a skill is a procedure","faq":null},{"slug":"payer-behavior-biggest-threat-to-practice-revenue","title":"Payer Behavior Is N
10ow the Biggest Threat to Practice Revenue â Here's What Small Practices Can Do About It","date":"2026-02-27T19:01:20.111Z","readingTime":"10 min read","excerpt":"New data from 120+ healthcare leaders shows payer denials and underpayments have become the #1 revenue risk for 2026. Here's what independent practices can do â and why AI agents are the answer.","category":"Healthcare","segments":["Healthcare"],"tags":["Agentic Healthcare","Denial Management","Payer Behavior","Revenue Cycle Management","Small Practice","Independent Practice","AI Agents"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/what-can-small-practices-do-when-payors-keep-denying-claims.png","heroImageAlt":"Illustration of revenue flowing from a medical practice while payer denials create a growing leak","faq":[{"question":"How many hours per week do practices spend on denial management?","answer":"Most healthcare organizations spend between 50 and 75 hours per week on denial-related work, according to the 2026 Adonis survey of 120+ healthcare leaders. For small practices, this represents a disproportionate share of total staff capacity, often consuming the equivalent of 1â2 full-time employees who could otherwise support patient care or revenue-generating activities."},{"question":"What percentage of revenue do small practices lose to payer denials?","answer":"Nearly half (47%) of organizations surveyed lose between 3% and 4% of net patient revenue to denied claims, underpayments, and timely filing limits. For a practice generating $2 million in annual revenue, that translates to $60,000â$80,000 per year â not from coding errors or poor care delivery, but from payer rule changes and missed appeal windows."},{"question":"Can AI replace billing staff in a medical practice?","answer":"No. AI agents for revenue cycle management handle surveillance, pattern recognition, and deadline tracking â not clinical or billing decisions. Every action that touches a patient or a payer goes through human review. The agent automates the high-volume, rules-based monitoring work that no human team can sustain across dozens of payers with constantly shifting rules, freeing billing staff to focus on complex cases and strategic decisions."},{"question":"What is the most important AI capability for healthcare RCM in 2026?","answer":"Automated denial follow-up and resolution ranks as the top priority, with 66% of healthcare leaders identifying it as \"very important\" for their 2026 strategy. The second priority is greater insight and transparency into payer activity, cited by 29% of leaders â reflecting a need for proactive intelligence rather than reactive claim processing (Adonis, 2026)."},{"question":"How do payer adjudication rule changes affect small practices differently than large health systems?","answer":"Large health systems absorb payer rule changes across hundreds of billing specialists and dedicated compliance teams. Small practices â typically operating with 2â3 billing staff â lack the capacity to monitor rule changes across dozens of payers simultaneously. When adjudication rules shift mid-quarter, small practices often discover the impact only after revenue has already been lost, because they don't have real-time payer performance data."}]},{"slug":"prior-auth-is-going-electronic-what-small-practices-need-to-know","title":"Prior Authorization Is Going Electronic â What Small Practices Should Know About CMS-0057","date":"2026-02-24T20:19:45.357Z","readingTime":"6 min read","excerpt":"CMS-0057-F is reshaping how prior authorization works â with faster decision timelines, specific denial reasons, and new FHIR APIs on the horizon. Here's what the shift means for small medical practices and how to prepare.","category":"Healthcare","segments":["Healthcare"],"tags":["Agentic Healthcare","Prior Authorization","CMS-0057","FHIR","Revenue Cycle Management","Interoperability","Small Practice"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/prior-auth-cms-0057.png","heroImageAlt":"Illustration of prior authorization workflow transitioning from manual fax to electronic FHIR API","faq":null},{"slug":"skillsbench-first-agent-skills-benchmark-validates-performance-gains","title":"The First Benchmark for Agent Skills Is Here â and the Results Validate Everything We've Been Building","date":"2026-02-23T08:52:15.249Z","readingTime":"8 min read","excerpt":"SkillsBench research proves curated Agent Skills raise AI performance by 16.2 percentage points. Healthcare gains hit +51.9pp. Here's what it means for production AI.","category":"Platform","segments":["Platform","Healthcare","Agent Skills"],"tags":["Agent Skills","SkillsBench","Research","Healthcare AI","MCP","myAgentSkills","Benchmarks"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/skillsbench-hero.png","heroImageAlt":"Abstract visualization of AI performance improvement through curated Agent Skills","faq":null},{"slug":"what-makes-a-great-agent-skill","title":"What Makes a Great AI Skill","date":"2026-02-20T00:00:00.000Z","readingTime":"3 min read","excerpt":"The anatomy of a well-crafted AI skill â from clear instructions to validation rules â and common mistakes to avoid when building your own.","category":"Agent Skills","segments":["Learn","Agent Skills"],"tags":["Learn","Skill Design","Best Practices","Agent Skills","AI Skills"],"author":{"name":"Agentman Academy","avatar":"/images/agentman-logo.png"},"heroImage":"/images/blog/agentman-academy-hero.png","heroImageAlt":"Agentman Academy â what makes a great AI skill","faq":null},{"slug":"24-months-in-production-the-missing-piece-finally-clicked","title":"24 Months in Production. The Missing Piece Finally Clicked.","date":"2025-12-11T00:00:00.000Z","readingTime":"8 min read","excerpt":"Every agent needs six components. Most platforms only give you three. Here's how Skills complete the agent ar
10chitecture.","category":"Platform","segments":["Platform","Agent Skills"],"tags":["Agentman","Skills","Agent Skills","AI Agents","Agent Architecture","AgentStudio","AgentForge","AgentLens","MCP"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/agentman-skills-library.png","heroImageAlt":"Agentman Skills Library showing skill-creator, brand-guidelines, internal-comms, mcp-builder, and other reusable skills","faq":null},{"slug":"their-agents-make-software-stickier-ours-make-it-optional","title":"Their Agents Make Software Stickier. Ours Make It Optional.","date":"2025-12-10T00:00:00.000Z","readingTime":"6 min read","excerpt":"Big Tech can't build agents that threaten their revenue. We can. Here's why platform independence matters for the future of AI agents.","category":"Platform","segments":["Platform"],"tags":["Agentman","Independence","AI Agents","Platform Neutrality","Big Tech","Digital Transformation","Vendor Lock-in"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/big-tech-vs-independent-agents.png","heroImageAlt":"Big Tech agents protect the stack vs Independent agents replace it - comparison of agent platform strategies","faq":null},{"slug":"12-months-of-agentic-chaos-zero-rewrites","title":"12 Months of Agentic Chaos. Zero Rewrites.","date":"2025-12-09T00:00:00.000Z","readingTime":"7 min read","excerpt":"From API tools to MCP servers. From solo builders to teams. Our architecture handled it allâwithout a single rewrite.","category":"Platform","segments":["Platform"],"tags":["Agentman","Architecture","AI Agents","AgentStudio","AgentForge","AgentLens","MCP","Multi-Agent"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/12-months-agentic-chaos.png","heroImageAlt":"12 months of agentic AI chaos - API tools to MCP servers, prompt engineering to context engineering, single agents to multi-agent architectures","faq":null},{"slug":"we-started-building-agents-before-anyone-believed","title":"We Started Building Agents Before Anyone Believed It Was Possible","date":"2025-12-08T00:00:00.000Z","readingTime":"6 min read","excerpt":"Two years ago, we bet on no-code agents when everyone said it couldn't be done. Here's what happened and why the market is finally catching up.","category":"Platform","segments":["Platform"],"tags":["Agentman","No-Code","AI Agents","Company Story","Agent Platform","AgentStudio","AgentForge","AgentLens"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/we-started-building-agents-before-anyone-believed.png","heroImageAlt":"Agentman platform overview showing AgentStudio, AgentForge, AgentLens, and AgentWatch - Ship AI Agents with confidence","faq":null},{"slug":"reverse-ngineering-latest-ChatGPT-memory-feature-and-building-your-own","title":"Reverse Engineering Latest ChatGPT Memory Feature (And Building Your Own)","date":"2025-04-22T19:10:00.000Z","readingTime":"10 min read","excerpt":"A deep dive into OpenAI's upgraded memory system with a step-by-step blueprint for building your own memory capabilities, no programming expertise required.","category":"Platform","segments":["Platform"],"tags":["AI","Memory Systems","Architecture","Personalization","Agentman","AI Agents","OpenAI","ChatGPT"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/chatgpt-memory-architecture-hero.png","heroImageAlt":"Person interacting with an AI assistant that visually demonstrates memory connections","faq":null},{"slug":"beyond-stateless-how-memory-makes-ai-agents-truly-intelligent","title":"Beyond Stateless: How Memory Makes AI Agents Truly Intelligent","date":"2025-04-22T19:00:00.000Z","readingTime":"6 min read","excerpt":"First in a series exploring how memory systems transform AI from question-answerers to persistent, personalized agents that can build relationships and complete complex tasks over time.","category":"Platform","segments":["Platform"],"tags":["AI","Memory Systems","Agents","LLM","Personalization"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/ai-memory-systems-hero.png","heroImageAlt":"Neural network with memory nodes highlighted, representing how memory systems transform AI agents","faq":null},{"slug":"mastering-vibe-coding-advanced-techniques","title":"Mastering Vibe Coding: 8 Advanced Techniques for AI-Assisted Development","date":"2025-04-07T00:00:00.000Z","readingTime":"9 min read","excerpt":"Learn 8 advanced vibe coding techniques that transform good results into exceptional ones. Based on real-world experience building Agentman's platform with AI assistance.","category":"Platform","segments":["Platform"],"tags":["Agentman","AI Development","Vibe Coding","Software Engineering","no-code building"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/8-advanced-vibe-coding-techniques-v4.png","heroImageAlt":"Visual representation of 8 advanced vibe coding techniques for AI-assisted development","faq":null},{"slug":"democratization-ai-agents-ecommerce","title":"The Democratization of AI Agents - Enterprise Power for Every Ecommerce Store","date":"2025-03-26T00:00:00.000Z","readingTime":"5 min read","excerpt":"Discover how Agentman is making enterprise-grade AI agents accessible to ecommerce stores of all sizes, transforming customer experiences and operational efficiency across platforms.","category":"Platform","segments":["Platform"],"tags":["AI Agents","Ecommerce","Democratization","Customer Experience","Operational Efficiency","Agentman","Multi-platform"],"author":{"name":"Gayatri Seetharaman","avatar":"/images/gayatri-seetharaman.jpeg"},"heroImage":"/images/blog/democratization-ai-agents-ecommerce-2.png","heroImageAlt":"Illustration of AI agents helping small ecommerce stores compete with enterprise retailers","faq":null},{"slug":"top-three-vibe-coding-tools","title":"Top Three Vibe Coding Tools to Go from 0 to Launch in Days Instead of Weeks","date":"2025-03-19T00:00:00.000Z","readingTime":"5 min read","excerpt":"Discover the AI-powered development tools that are revolutionizing software creation, allowing teams to build and deploy projects in days rather than weeks.","category":"Platform","segments":["Platform"],"tags":["AI","Development","Vibe Coding","Productivity","Tools"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/top-3-vibe-coding-tools-hero-2.png","heroImageAlt":"Top 3 Vibe Coding tools: Illustration of AI-powered development tools accelerating software creation","faq":null},{"slug":"what-the-hell-is-vibe-coding","title":"What the Hell is Vibe Coding?","date":"2025-03-19T00:00:00.000Z","readingTime":"5 min read","excerpt":"Discover how vibe coding is transforming software development by allowing developers to describe what they want in natural language while AI handles the implementation details.","category":"Platform","segments":["Platform"],"tags":["AI","Development","Vibe Coding","Software Engineering","Generative AI"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/vibe-coding-hero-4.png","heroImageAlt":"Vibe coding:Illustration of a developer collaborating with AI to write code","faq":null},{"slug":"pricing-strategies-agentic-startups","title":"Pricing Strategies Among Agentic Startups - A Competitive Analysis","date":"2025-03-09T00:00:00.000Z","readingTime":"10 min read","excerpt":"Explore how AI agent platforms are pricing their offerings, from freemium to enterprise models, and discover how Agentman's approach balances flexibility with predictability.","category":"Platform","segments":["Platform","PE"],"tags":["AI Agents","Pricing Strategy","S
10aaS Business","Competitive Analysis","Agentman","Startup Growth","AI Monetization"],"author":{"name":"Gayatri Seetharaman","avatar":"/images/gayatri-seetharaman.jpeg"},"heroImage":"/images/blog/pricing-strategies-agentic-startups.png","heroImageAlt":"Visual representation of different pricing tiers and strategies for AI agent platforms","faq":null},{"slug":"why-your-business-needs-AI-agents-not-another-chatbot","title":"Why Your Business Needs AI Agents, Not Just Another Chatbot","date":"2025-02-17T00:00:00.000Z","readingTime":"5 min read","excerpt":"Unlock the full potential of AI agents in transforming your business operations, customer experience, and revenue growth.","category":"Platform","segments":["Platform"],"tags":["AI Agents","Digital Transformation","Business Innovation","Customer Experience","Operational Efficiency","Revenue Growth"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/understanding-ai-agents-automation.png","heroImageAlt":"Visual comparison between limited chatbot capabilities and sophisticated AI agent interactions","faq":null},{"slug":"complete-guide-to-ai-agent-pricing-models-2025","title":"The Complete Guide to AI Agent Pricing Models in 2025","date":"2025-01-20T00:00:00.000Z","readingTime":"5 min read","excerpt":"An in-depth exploration of the four core pricing models for AI agents, from per-execution to usage-based pricing, and how to choose the right model for your business.","category":"Platform","segments":["Platform"],"tags":["AI Agents","Pricing Models","Business Strategy","Enterprise AI","SMB Solutions","Cost Analysis","ROI"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/the-complete-guide-to-ai-agent-pricing-models-in-2025-sm.png","heroImageAlt":"Visual comparison of different AI agent pricing models and their business impact","faq":null},{"slug":"introducing-guardian-agents-the-future-of-agent-assistance","title":"Introducing Guardian Agents (Ambient Agents): The Future of Agent Assistance","date":"2025-01-15T17:45:24.000Z","readingTime":"5 min read","excerpt":"Discover how Guardian Agents (Ambient Agents) are revolutionizing AI assistance by working proactively in the background, protecting your time and attention while taking care of what matters.","category":"Platform","segments":["Platform"],"tags":["Guardian Agents","Ambient Agents","AI Assistance","Business Automation","Productivity","Future of Work","AI Innovation"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/guardian-agents-ambient-agents-cover.webp","heroImageAlt":"Conceptual illustration of Guardian Agents working seamlessly in the background while humans focus on strategic tasks","faq":null},{"slug":"7-trends-shaping-ai-agents-future","title":"From No-Code Builders to Agent-First Apps - 7 Trends Shaping AI Agents","date":"2024-10-13T00:00:00.000Z","readingTime":"6 min read","excerpt":"From no-code builders to generative UI, discover how AI agents are revolutionizing the way businesses operate and innovate.","category":"Platform","segments":["Platform"],"tags":["AI Agents","No-Code Development","AI Trends","Future of AI","Enterprise AI","Agent-First Apps","Business Innovation"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/7-Trends-Shaping-the-Future-of-AI-Agents-Future.jpg","heroImageAlt":"Visual representation of AI agent evolution showing the progression from code-based to no-code platforms","faq":null},{"slug":"the-rise-of-the-ai-agent-product-manager-and-ai-agent-engineer","title":"The Rise of the AI Agent Product Manager and AI Agent Engineer","date":"2024-10-02T00:00:00.000Z","readingTime":"9 min read","excerpt":"Discover the emerging roles of AI Agent Product Manager and AI Agent Engineer, and how they're shaping the future of business operations and innovation.","category":"Platform","segments":["Platform"],"tags":["Generative Agents","Language Agents","Generative AI Use Cases","AI Careers","Product Management","AI Engineering","Business Innovation","AI Strategy"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/Agent-PM-and-Agent-Engineer.png","heroImageAlt":"Primary responsibilities of AI Agent Product Manager and AI Agent Engineer","faq":null},{"slug":"strategic-framework-llm-applications","title":"A Strategic Framework for Building LLM Applications","date":"2024-02-14T00:00:00.000Z","readingTime":"7 min read","excerpt":"Discover a comprehensive framework for building LLM applications, from off-the-shelf solutions to custom-trained models.","category":"Platform","segments":["Platform"],"tags":["LLM Applications","Fine Tuning","Functions And Agents","Strategic Frameworks","Retrieval Augmented","AI Development","Machine Learning","ChatGPT","Enterprise AI","Generative AI","AI Strategy"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/Strategic-framework-for-Building-Generative-AI-in-Your-Business.png","heroImageAlt":"Visual representation of different approaches to building LLM applications","faq":null},{"slug":"why-your-business-needs-to-blend-human-expertise-and-ch
10atGPT","title":"Why Your Business Needs to Blend Human Expertise and ChatGPT","date":"2023-06-07T00:00:00.000Z","readingTime":"6 min read","excerpt":"Discover how combining human expertise with ChatGPT's capabilities can transform your business operations and drive innovation.","category":"Platform","segments":["Platform"],"tags":["GPT 4","Reasoning","Knowledge","Large Language Models","Applications Of AI","Business Strategy","AI Innovation"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/human-expertise-and-chatgpt.webp","heroImageAlt":"Visual representation of human expertise working alongside AI","faq":null},{"slug":"introduction-to-llms-and-applications","title":"Introduction to LLMs and their Applications","date":"2023-05-05T00:00:00.000Z","readingTime":"5 min read","excerpt":"A comprehensive introduction to Large Language Models (LLMs), their evolution, key concepts, and practical applications for full-stack engineers.","category":"Platform","segments":["Platform"],"tags":["Large Language Models","Full Stack Development","Machine Learning AI","Product Development","AI Applications","GPT","Natural Language Processing"],"author":{"name":"Prasad Thammineni","avatar":"/images/prasad-thammineni.jpeg"},"heroImage":"/images/blog/introduction-to-llms.webp","heroImageAlt":"Introduction to Large Language Models and their Applications","faq":null}]},"__N_SSG":true},"page":"/blog","query":{},"buildId":"W-uwlKWK8YmOc7QuKgaOl","isFallback":false,"gsp":true,"scriptLoader":[]}</script>
10</body></html>
Line numbers count LF bytes from the start of the resource, as the search results do. Vendor segments are library code the classifier recognised; they are stored but not indexed. Bytes are shown as Latin1 characters, one per byte.