PageSourceSearch

https://www.opificiolamantinianonimi.com/en/blog/ai-in-business-18-month-review-2027-compass

html opificiolamantinianonimi.com collected 2026-10-06 12:39:13 UTC 80,843 bytes, 65 lines download raw bytes

1<!DOCTYPE html><html lang="en"><head><meta charSet="utf-8" data-next-head=""/><meta name="viewport" content="width=device-width" data-next-head=""/><title data-next-head="">AI in business: 18-month review and compass for 2027 | Opificio Bites</title><meta name="description" content="Eighteen months of generative AI applied to Italian SMEs. What worked, what was hidden cost, where to invest in the next 12 months. The long-form closing of the summer — to re-read at Christmas to see if we got it right." data-next-head=""/><link rel="canonical" href="https://www.opificiolamantinianonimi.com/blog/ai-in-business-18-month-review-2027-compass" data-next-head=""/><link rel="alternate" hrefLang="it" href="https://www.opificiolamantinianonimi.com/blog/ai-in-business-18-month-review-2027-compass" data-next-head=""/><link rel="alternate" hrefLang="en" href="https://www.opificiolamantinianonimi.com/en/blog/ai-in-business-18-month-review-2027-compass" data-next-head=""/><link rel="alternate" hrefLang="x-default" href="https://www.opificiolamantinianonimi.com/blog/ai-in-business-18-month-review-2027-compass" data-next-head=""/><meta property="og:site_name" content="Opificio Lamantini Anonimi" data-next-head=""/><meta property="og:url" content="https://www.opificiolamantinianonimi.com/blog/ai-in-business-18-month-review-2027-compass" data-next-head=""/><meta property="og:type" content="article" data-next-head=""/><meta property="og:title" content="AI in business: 18-month review and compass for 2027 | Opificio Bites" data-next-head=""/><meta property="og:description" content="Eighteen months of generative AI applied to Italian SMEs. What worked, what was hidden cost, where to invest in the next 12 months. The long-form closing of the summer — to re-read at Christmas to see if we got it right." data-next-head=""/><meta property="og:image" content="https://www.opificiolamantinianonimi.com/images/general.jpg" data-next-head=""/><meta property="og:locale" content="en_US" data-next-head=""/><meta property="og:locale:alternate" content="it_IT" data-next-head=""/><meta name="twitter:card" content="summary_large_image" data-next-head=""/><meta name="twitter:site" content="@opaboratorio" data-next-head=""/><meta name="twitter:title" content="AI in business: 18-month review and compass for 2027 | Opificio Bites" data-next-head=""/><meta name="twitter:description" content="Eighteen months of generative AI applied to Italian SMEs. What worked, what was hidden cost, where to invest in the next 12 months. The long-form closing of the summer — to re-read at Christmas to see if we got it right." data-next-head=""/><meta name="twitter:url" content="https://www.opificiolamantinianonimi.com/blog/ai-in-business-18-month-review-2027-compass" data-next-head=""/><meta name="twitter:image" content="https://www.opificiolamantinianonimi.com/images/general.jpg" data-next-head=""/><link rel="apple-touch-icon" sizes="57x57" href="/favicon/apple-icon-57x57.png"/><link rel="apple-touch-icon" sizes="60x60" href="/favicon/apple-icon-60x60.png"/><link rel="apple-touch-icon" sizes="72x72" href="/favicon/apple-icon-72x72.png"/><link rel="apple-touch-icon" sizes="76x76" href="/favicon/apple-icon-76x76.png"/><link rel="apple-touch-icon" sizes="114x114" href="/favicon/apple-icon-114x114.png"/><link rel="apple-touch-icon" sizes="120x120" href="/favicon/apple-icon-120x120.png"/><link rel="apple-touch-icon" sizes="144x144" href="/favicon/apple-icon-144x144.png"/><link rel="apple-touch-icon" sizes="152x152" href="/favicon/apple-icon-152x152.png"/><link rel="apple-touch-icon" sizes="180x180" href="/favicon/apple-icon-180x180.png"/><link rel="icon" type="image/png" sizes="192x192" href="/favicon/android-icon-192x192.png"/><link rel="icon" type="image/png" sizes="32x32" href="/favicon/favicon-32x32.png"/><link rel="icon" type="image/png" sizes="96x96" href="/favicon/favicon-96x96.png"/><link rel="icon" type="image/png" sizes="16x16" href="/favicon/favicon-16x16.png"/><link rel="manifest" href="/favicon/manifest.json"/><meta name="msapplication-TileColor" content="#293038"/><meta name="msapplication-TileImage" content="/favicon/ms-icon-144x144.png"/><meta name="theme-color" content="#293038"/>
1<link rel="preconnect" href="https://use.typekit.net"/><link rel="preconnect" href="https://cdn.segment.com"/><link rel="preconnect" href="https://www.googletagmanager.com"/><link rel="dns-prefetch" href="https://connect.facebook.net"/><link rel="dns-prefetch" href="https://snap.licdn.com"/><link rel="dns-prefetch" href="https://opificio.opificiolamantinianonimi.com"/><meta name="google-site-verification" content="zrCltmwyD-2lGr4SEximauBBWeeyuQ9hRcya1y1H6d8"/><link rel="preload" href="/_next/static/chunks/04d35m1pkllvt.css" as="style"/><link rel="preload" href="/_next/static/chunks/2141v81xzpnw7.css" as="style"/>
1<script type="application/ld+json" data-next-head="">{"@context":"https://schema.org","@type":"BlogPosting","mainEntityOfPage":{"@type":"WebPage","@id":"https://www.opificiolamantinianonimi.com/blog/ai-in-business-18-month-review-2027-compass"},"headline":"AI in business: 18-month review and compass for 2027","description":"Eighteen months of generative AI applied to Italian SMEs. What worked, what was hidden cost, where to invest in the next 12 months. The long-form closing of the summer — to re-read at Christmas to see if we got it right.","image":"undefined","author":[{"@type":"Person","name":"Luca Manuele Simonato","url":"https://www.opificiolamantinianonimi.com/luca","@id":"https://www.opificiolamantinianonimi.com/luca"},{"@type":"Person","name":"Manuel Rigo","url":"https://www.opificiolamantinianonimi.com/manuel","@id":"https://www.opificiolamantinianonimi.com/manuel"},{"@type":"Person","name":"Enrico Quaglio","url":"https://www.opificiolamantinianonimi.com/enrico","@id":"https://www.opificiolamantinianonimi.com/enrico"}],"publisher":{"@type":"Organization","@id":"https://www.opificiolamantinianonimi.com","name":"Opificio Lamantini Anonimi","logo":{"@type":"ImageObject","url":"https://www.opificiolamantinianonimi.com/images/OLA-logo.png"}},"datePublished":"2026-08-13","dateModified":"2026-05-18T10:38:10.405Z","wordCount":1800}</script>
1<script type="application/ld+json" data-next-head="">{"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://www.opificiolamantinianonimi.com/"},{"@type":"ListItem","position":2,"name":"Blog","item":"https://www.opificiolamantinianonimi.com/blog"},{"@type":"ListItem","position":3,"name":"AI in business: 18-month review and compass for 2027","item":"https://www.opificiolamantinianonimi.com/blog/ai-in-business-18-month-review-2027-compass"}]}</script>
1<noscript><img height="1" width="1" style="display:none" src="https://www.facebook.com/tr?id=655181521892520&amp;ev=PageView&amp;noscript=1"/></noscript><noscript><img height="1" width="1" style="display:none" alt="" src="https://px.ads.linkedin.com/collect/?pid=1787433&amp;fmt=gif"/></noscript><link rel="stylesheet" href="/_next/static/chunks/04d35m1pkllvt.css" data-n-g=""/><link rel="stylesheet" href="/_next/static/chunks/2141v81xzpnw7.css" data-n-g=""/><noscript data-n-css=""></noscript>
1<script src="/_next/static/chunks/3b9znn8o6ck8y.js" defer=""></script>
1<script src="/_next/static/chunks/0on3o95lq477w.js" defer=""></script>
1<script src="/_next/static/chunks/2t8_o23cvyy1k.js" defer=""></script>
1<script src="/_next/static/chunks/2_nu27t0pqied.js" defer=""></script>
1<script src="/_next/static/chunks/turbopack-37lvz01p1r7dh.js" defer=""></script>
1<script src="/_next/static/chunks/0g4su_jo8onmc.js" defer=""></script>
1<script src="/_next/static/chunks/1mwusd3bc7gfd.js" defer=""></script>
1<script src="/_next/static/chunks/0blx_g6is2qop.js" defer=""></script>
1<script src="/_next/static/chunks/4089fflp147oc.js" defer=""></script>
1<script src="/_next/static/chunks/1i1v4rmi0dmc2.js" defer=""></script>
1<script src="/_next/static/chunks/0dtrnui_14w9a.js" defer=""></script>
1<script src="/_next/static/chunks/17khgdi889t07.js" defer=""></script>
1<script src="/_next/static/chunks/0v1h44hup3d_k.js" defer=""></script>
1<script src="/_next/static/chunks/turbopack-3hl4kaxhkzso1.js" defer=""></script>
1<script src="/_next/static/Wk-8BYbl35XFN6ORArdsv/_buildManifest.js" defer=""></script>
1<script src="/_next/static/Wk-8BYbl35XFN6ORArdsv/_ssgManifest.js" defer=""></script>
1<script src="/_next/static/Wk-8BYbl35XFN6ORArdsv/_clientMiddlewareManifest.js" defer=""></script>
1</head><body><div id="__next"><div class="main-layout"><header class="fixed w-full h-[80px] px-6 box-border z-[1000] flex items-center justify-between lg:hidden transition-all"><div><a href="/en"><img alt="Opificio Lamantini Anonimi" loading="lazy" width="45" height="45" decoding="async" data-nimg="1" style="color:transparent" srcSet="/_next/image?url=%2Fimages%2FOLA-lamantino-w.png&amp;w=48&amp;q=75 1x, /_next/image?url=%2Fimages%2FOLA-lamantino-w.png&amp;w=96&amp;q=75 2x" src="/_next/image?url=%2Fimages%2FOLA-lamantino-w.png&amp;w=96&amp;q=75"/></a></div><div class="relative m-0 top-0 left-0 flex items-center justify-between text-center font-bold text-xl h-[12px] overflow-hidden text-white ">menu</div></header><div class="fixed top-10 left-10 hidden h-[50px] w-[70px] lg:flex items-center justify-center z-[1000] "><a href="/en"><img alt="Opificio Lamantini Anonimi" loading="lazy" width="60" height="50" decoding="async" data-nimg="1" style="color:transparent" srcSet="/_next/image?url=%2Fimages%2FOLA-lamantino-w.png&amp;w=64&amp;q=75 1x, /_next/image?url=%2Fimages%2FOLA-lamantino-w.png&amp;w=128&amp;q=75 2x" src="/_next/image?url=%2Fimages%2FOLA-lamantino-w.png&amp;w=128&amp;q=75"/></a></div><div class="hidden lg:flex fixed top-10 right-10 h-[50px] w-[70px] items-center justify-center z-[1000] hover:opacity-50 transition-all  "><div class="m-0 top-0 left-0 flex items-center justify-between text-center font-bold text-2xl h-[12px] overflow-hidden text-white">menu</div></div><div class="pb-12 lg:pb-20 bg-white"><div class="relative w-full mx-auto px-6 md:px-20 lg:px-32 box-border max-w-[1200px] max-w-[1200px] pt-14"><div class="flex flex-col md:flex-row gap-5 lg:gap-10"><div><div class="font-bold text-red-new uppercase pb-2 leading-none flex items-center gap-3"><span>13 ago 2026</span><span class="text-blue-new/50 normal-case font-normal text-sm">9<!-- --> min <!-- -->read</span></div><h1 class="font-bold text-blue-new text-4xl">AI in business: 18-month review and compass for 2027</h1><h2 class="font-bold text-blue-new text-2xl pt-4 leading-tight">Eighteen months of generative AI applied to Italian SMEs. What worked, what was hidden cost, where to invest in the next 12 months. The long-form closing of the summer — to re-read at Christmas to see if we got it right.</h2></div></div><div class="pt-10 pb-10 px-6 lg:px-10 bg-red-new my-8 lg:my-14"><div class="font-bold lg:text-2xl text-white border-l-2 border-white pl-4 py-2"><p>&quot;2024 was the year of the promise. 2025 was the year of disappointment. 2026 is the year of discipline. Those who understand this trajectory survive the next 18 months without ruining themselves.&quot;</p></div></div><div class="article-content whitespace-break-spaces leading-[1.5] lg:text-lg"><p>In early 2025, a boardroom of an Italian SME in the Northeast. Four people: CEO, CFO, IT head, marketing. The agenda item is &quot;define the company&#x27;s AI strategy&quot;. On the table is a stack of printed newsletters, a consulting firm&#x27;s 80-page report, two quotes from different vendors, and the weekend edition of Sole 24 Ore with a full-page headline on generative AI.</p>
2<p>The meeting lasts three hours. They come out with a decision: &quot;evaluate over the next six months&quot;. In June 2025 a plan is presented. In September 2025 it&#x27;s postponed. In January 2026 it&#x27;s resurrected. In May 2026 a budget is approved. In September 2026 — now — implementation will begin.</p>
3<p>Seventeen months.</p>
4<p>Yes, the company is a real case. No, we can&#x27;t name it. But it&#x27;s exactly <em>the average Italian company</em> — among the SMEs that did something on AI in 2024-2026.</p>
5<p>This piece is a long-form review. Eighteen months of generative AI applied to Italian companies, distilled into what we&#x27;ve seen work, what turned out to be hidden cost, and where it&#x27;s worth investing in the next 12 months. It&#x27;s deliberately long. It&#x27;s deliberately frank. It&#x27;s deliberately written <strong>to be re-readable at Christmas</strong> — because many things said here might already be outdated, and those that won&#x27;t be will be our shared compass for 2027.</p>
6<h2>Three phases, one trajectory</h2>
7<p>In 18 months, AI in business went through three clear phases.</p>
8<p><strong>Phase 1 — The promise (January 2024 — June 2025).</strong> Everyone wants to do AI. Nobody knows exactly what to do with it. Budgets inflate, quotes proliferate, &quot;ChatGPT for managers&quot; courses multiply. AI strategy is a statement of intent: <em>&quot;we want to be AI-first&quot;</em>. Underneath, little concrete.</p>
9<p><strong>Phase 2 — The disappointment (July 2025 — December 2025).</strong> First pilots go badly or go <em>less well than expected</em>. Companies discover that integrating an LLM into their operations is more complex than the demo suggested. CMOs who pushed AI projects come under pressure from the CFO. Some initiatives get frozen, others reduced. <em>&quot;AI deception&quot;</em> becomes common talk.</p>
10<p><strong>Phase 3 — Discipline (January 2026 — today).</strong> Companies that survived phase 2 stopped asking &quot;should we do AI?&quot; and started asking &quot;<em>for which specific problem is AI the best solution?</em>&quot;. Pilots become smaller, more targeted, more measurable. The vocabulary refines: <em>generative AI</em>, <em>agentic AI</em>, <em>RAG</em>, <em>context engineering</em>, <em>multimodal</em> are distinguished. Decisions are made with discipline.</p>
11<p>The three phases are not over. They coexist. Every new company entering the AI market lives the three phases <em>in compressed scale</em>: from promise to discipline in a few months if lucky, in two years if less attentive. But the trajectory is clear, and it&#x27;s that one.</p>
12<h2>What really worked (in Italian SMEs)</h2>
13<p>We give you the real cases, anonymized but recognizable as patterns. They are the AI applications where we saw <strong>verified ROI</strong> in Italian SMEs.</p>
14<p><strong>1. AI-assisted (not replaced) customer service.</strong> For companies with recurring weekly ticket volumes (services, e-commerce, software), an AI assistant that <em>prepares</em> human responses (doesn&#x27;t send them directly) reduces average handling time by 40-60%. It works because it preserves human responsibility, but eliminates repetitive typing. Investment: medium. Time to return: 3-6 months.</p>
15<p><strong>2. Internal data research and analysis.</strong> Companies with lots of internal documentation (legal, consulting, manufacturers with technical manuals) found great value in RAG systems (see <a href="/blog/hardwords-rag">the hardword piece on RAG</a>) that allow querying the archive in natural language. It works because it solves a real problem — <em>&quot;I have a specific question and the answer is in one of our 4,000 documents&quot;</em>. Investment: medium. Time to return: 6-9 months.</p>
16<p><strong>3. Creative variant generation.</strong> Marketing teams managing many campaigns used AI to generate copy and creative ad variants, accelerating testing. Doesn&#x27;t replace the creative. <em>Amplifies them.</em> Works very well for ambient marketing, performance marketing, social ads. Investment: low. Time to return: 1-3 months.</p>
17<p><strong>4. Accelerated software development (with rules).</strong> As written in the <a href="/blog/vibe-coding-in-production">July 23 vibe coding piece</a>, internal development teams got +50-80% release speed. <em>With governance.</em> Investment: medium-high (in training and processes). Time to return: 6 months.</p>
18<p><strong>5. Complex document audit.</strong> Professional firms and consultancies are using multimodal AI for first passes on balance sheets, contracts, technical sheets. Human review remains indispensable, but initial &quot;decoding&quot; time collapses. Works very well for accounting firms, mid-sized law firms, tax consultants. Investment: low. Time to return: 3 months.</p>
19<h2>What was hidden cost</h2>
20<p>
20Now the part less willingly told. Three things that, in many SMEs, turned out to be expensive without anyone noticing first.</p>
21<p><strong>1. Time spent &quot;evaluating vendors&quot;.</strong> Italian companies above a certain size spent dozens of half-days of the CFO, IT head, marketing, in meetings with AI vendors in 2024-2025. The vast majority of these meetings produced no decisions. Submerged opportunity cost: easily 30-50 thousand euros of managerial time per average company. On a thousand Italian SMEs, we&#x27;re talking about <strong>30-50 million euros a year</strong> of managerial time burned on evaluations. Few notice because it&#x27;s not in any cost center.</p>
22<p><strong>2. Overlapping &quot;tool stacks&quot;.</strong> Many companies today pay simultaneously: ChatGPT Enterprise licenses, Claude Team licenses, Copilot licenses for Microsoft 365, and one or two vertical tools (for chatbots, for analysis, for coding). Aggregate cost: often 30-60 thousand euros a year for a 50-employee company. Effective usage: ridiculously low. Rationalizing the tool stack is one of the most profitable exercises of Q4 2026.</p>
23<p><strong>3. Late-managed compliance.</strong> As written in <a href="/blog/ai-act-phase-2">the June 25 piece on AI Act phase 2</a>, many companies discovered <em>after</em> putting AI systems into production that they had to document, evaluate, and supervise. The cost of doing it <em>after</em> is triple compared to doing it <em>before</em>. Italian SMEs that structured AI governance practices already in 2025 are saving money and time in 2026.</p>
24<h2>The compass for 2027</h2>
25<p>Let&#x27;s come to the prediction part. Four trends we see consolidating, and which you should budget for in the next 12 months.</p>
26<h3>Trend 1 — AI becomes &quot;vertical&quot;, not &quot;horizontal&quot;</h3>
27<p>For the first two years of generative AI in business, the dominant tools were horizontal: ChatGPT, Claude, Gemini, Copilot. Tools that do <em>many things mediocrely well</em>.</p>
28<p>In 2027 value shifts to <em>vertical</em> tools. AI systems built specifically for: accounting, legal management, manufacturing, healthcare, consulting, education. They work <em>much better</em> for the specific case — because they&#x27;re designed around it — and in return <em>they only work for that</em>.</p>
29<p><strong>What to do:</strong> evaluate what you really need, and look for the <em>vertical</em> vendor that solves it. Stop thinking &quot;ChatGPT is the universal base&quot;.</p>
30<h3>Trend 2 — AI agents start replacing workflows (not people)</h3>
31<p>Agentic AI — systems that execute chains of actions autonomously — is moving from promise to operability in the next 12 months. But the way it replaces people is different from what everyone thought.</p>
32<p>It doesn&#x27;t replace customer service. It replaces <em>the customer service workflow</em> — the three different tools the operator uses to handle a case. The operator remains, but their work environment simplifies radically.</p>
33<p><strong>What to do:</strong> identify the most <em>friction-filled</em> workflows in your company (those requiring passages between many tools, many hands, many waits). Agents will enter there. Not elsewhere.</p>
34<h3>Trend 3 — &quot;AI-native&quot; become a new segment</h3>
35<p>In 2027 we&#x27;ll see companies emerge that are <strong>AI-native from birth</strong>: organizations built with the assumption that AI is a constitutive part of the process, not an addition. It&#x27;s what we&#x27;re trying to do with <a href="https://opificio.ai">OpificioAI</a>, Tuken, BP, BI, and Blogger. We don&#x27;t do it for fashion — we do it because we understood, in these 18 months, that AI-native companies <em>operate in a structurally different way</em>. They have smaller teams, faster decisions, products that scale with almost zero marginal cost.</p>
36<p><strong>What to do:</strong> if you&#x27;re planning a new business unit, evaluate building it AI-native from the start. On existing products it&#x27;s difficult to retrofit. On new ones it&#x27;s the right move.</p>
37<h3>Trend 4 — Trust becomes the real asset</h3>
38<p>In 2026 we saw a strange phenomenon: some technically <em>worse</em> AI vendors won contracts against technically <em>better</em> vendors. Why? Because they knew how to build <em>trust</em>. Data transparency, cost clarity, ready compliance documentation, open audit processes.</p>
39<p>In 2027 this trend will amplify. Companies will increasingly buy AI from vendors who can <strong>demonstrate</strong>, not just <strong>declare</strong>, to be reliable. Trust becomes part of the product.</p>
40<p><strong>What to do:</strong> if you&#x27;re an AI vendor, invest in transparency and documentation now. If you&#x27;re a customer, choose vendors based on transparency, not just performance.</p>
41<h2>Five operational tips for the next quarter</h2>
42<p>To close, five concrete actions for Q4 2026. All doable in 30-90 days.</p>
43<ol>
44<li>
45<p><strong>AI tool stack audit.</strong> How many tools do you pay for? How used are they? What can you consolidate? Typically: -30% annual costs, +20% clarity.</p>
46</li>
47<li>
48<p><strong>Audit of AI systems in production.</strong> How many are in &quot;high risk&quot; tier per the AI Act? For each, is there a compliance plan? Involve legal. It&#x27;s done in half a day, avoids millions of problems.</p>
49</li>
50<li>
51<p><strong>Define the product metric, not the agent.</strong> For each AI system in the company, what&#x27;s the <em>business</em> metric telling you if it&#x27;s working? If you don&#x27;t have it, build it. A clear metric beats a thousand engineering dashboards.</p>
52</li>
53<li>
54<p><strong>Invest in governance before the next tool.</strong> House rules (see <a href="/blog/ai-agents-in-production-4-recurring-mistakes">the May 28 piece on the 4 recurring mistakes</a>
54) are worth more than a new model, a new agent, a new license. Spend there.</p>
55</li>
56<li>
57<p><strong>Start a single vertical pilot, not a horizontal one.</strong> See trend 1. Identify the most specific problem in your company where AI can help, and put three months on it. More focus, more ROI, more learning.</p>
58</li>
59</ol>
60<h2>The final note for those reading us</h2>
61<p>This is the last piece of the May-August 2026 editorial plan. Twenty-six articles, written two a week, without skipping any. A promise we made you in May when we reopened the blog after seven months of silence, and that we kept.</p>
62<p>In September we&#x27;ll be back with a new editorial block. Topics, cadence, tone — we&#x27;ll define them in the coming weeks. If you have suggestions, write to us. This blog isn&#x27;t a monologue, it&#x27;s a conversation.</p>
63<p>Happy Ferragosto. Good return. See you in September, fresh, with some more ideas than now. And maybe — with a little discipline and a good dose of curiosity — one more piece of AI that really works in our SMEs.</p>
64<hr/>
65<p><strong>Do you feel halfway between the disappointment phase and the discipline phase?</strong> Under <a href="/intelligenza-artificiale-per-il-business">artificial intelligence for business</a> and with <a href="https://opificio.ai">OpificioAI</a> we help Italian SMEs choose and put into production the AI that makes a difference. <a href="/contatti">Let&#x27;s talk</a> in September, when you&#x27;re really ready.</p></div></div></div><div id="form" class="landing-form-container"><div class="relative w-full mx-auto px-6 md:px-20 lg:px-32 box-border max-w-[1200px]"><div class="landing-form-content"><div class="landing-form-left"><p class="landing-form-title"></p><div class="landing-form-description"></div></div><div class="landing-form-right"><div class="landing-form-paragraph">Che sia un’idea, una curiosità, una sfida da affrontare, per noi non è mai “solo un contatto”.<br /><br />È l’inizio di una conversazione, magari davanti a un caffè, reale o virtuale che sia.<br /><br />Compila il form qui sotto e raccontaci cosa ti passa per la testa. <br /><br />Promesso: niente automatismi, solo lamantini veri (con tastiera e cervello ben accesi).</div><form id="landing-form"><div class="landing-form"><div class="input-container half"><label>Nome *</label><input type="text" name="firstName"/></div><div class="input-container half"><label>Cognome *</label><input type="text" name="lastName"/></div><div class="input-container half"><label>Azienda</label><input type="text" required="" name="company"/></div><div class="input-container half"><label>Email *</label><input type="email" required="" name="email"/></div><div class="input-container input-consent"><label for="privacyConsent"><input id="privacyConsent" type="checkbox" required="" name="privacyConsent"/><span>Ho letto e accetto i termini e l’<a href="https://www.iubenda.com/privacy-policy/61956652" target="_blank" rel="noopener noreferrer">informativa sulla privacy</a>, e acconsento al trattamento dei miei dati per essere ricontattato. *</span></label></div><button class="landing-button mouse-cursor-simple form-submit" type="submit">Prenota</button></div></form></div></div></div></div><div class="py-14 lg:py-20 bg-red-new"><div class="relative w-full mx-auto px-6 md:px-20 lg:px-32 box-border max-w-[1200px] max-w-[1200px]"><div class="text-blue-new font-bold text-xl">Other Lamantini Bites</div><div class="py-4 lg:py-10 lg:grid lg:grid-cols-2 gap-20"><a class="block group relative py-3 lg:p-0" href="/en/blog/seasonal-brands"><p class="text-lg font-bold text-lightred-new uppercase lg:group-hover:text-blue-new">10 ago 2026</p><p class="text-2xl text-white font-bold pt-2 lg:group-hover:!text-blue-new">Seasonal brands: who knows how to use the emotional calendar (and who fakes it)</p><span class="absolute bottom-0 left-0 h-1 w-5 bg-white group-last:hidden lg:hidden"></span></a><a class="block group relative py-3 lg:p-0" href="/en/blog/hardwords-multimodal"><p class="text-lg font-bold text-lightred-new uppercase lg:group-hover:text-blue-new">06 ago 2026</p><p class="text-2xl text-white font-bold pt-2 lg:group-hover:!text-blue-new">#HardWords: Multimodal</p><span class="absolute bottom-0 left-0 h-1 w-5 bg-white group-last:hidden lg:hidden"></span></a><a class="block group relative py-3 lg:p-0" href="/en/blog/september-pitch-3-things-to-fix"><p class="text-lg font-bold text-lightred-new uppercase lg:group-hover:text-blue-new">03 ago 2026</p><p class="text-2xl text-white font-bold pt-2 lg:group-hover:!text-blue-new">The September pitch: 3 things to fix in August if you have a round in sight</p><span class="absolute bottom-0 left-0 h-1 w-5 bg-white group-last:hidden lg:hidden"></span></a><a class="block group relative py-3 lg:p-0" href="/en/blog/the-end-of-easy-seo-traffic"><p class="text-lg font-bold text-lightred-new uppercase lg:group-hover:text-blue-new">30 lug 2026</p><p class="text-2xl text-white font-bold pt-2 lg:group-hover:!text-blue-new">The end of &#x27;easy&#x27; SEO traffic: how to rebuild a realistic-driven plan</p><span class="absolute bottom-0 left-0 h-1 w-5 bg-white group-last:hidden lg:hidden"></span></a></div><a class="group w-full md:w-280 md:max-w-[280px] min-w-[180px] md:w-auto md:max-w-[280px] h-14 px-8 mx-auto md:mx-0 flex justify-center text-lg items-center gap-2 rounded-full font-medium cursor-pointer bg-[transparent] border border-blue-new transition-all !border-white !text-white" type="button" to="/blog" href="/en/blog">All articles</a></div></div><footer><div class="relative w-full mx-auto px-6 md:px-20 lg:px-32 box-border max-w-[1200px]"><div class="footer-top"><div class="footer-top-left"><p class="footer-top-title"><strong>Opificio Lamantini Anonimi s.r.l.</strong></p><p class="footer-top-subtitle">Studio di branding e innovazione digitale</p></div><div class="footer-top-right"><ul class="footer-sitemap"><li><a class="mouse-cursor-simple" href="/en/about">lamantini</a></li><li><a class="mouse-cursor-simple" href="/en/casi-studio">Case studies</a></li><li><a class="mouse-cursor-simple" href="/en/servizi">Services</a></li><li><a class="mouse-cursor-simple" href="/en/blog">Blog</a></li><li><a class="mouse-cursor-simple" href="/en/contatti">Contacts</a></li></ul></div></div><div class="footer-bottom"><div class="footer-bottom-left"><a target="_blank" rel="noopener noreferrer" class="footer-contacts" href="https://goo.gl/maps/88LBDucRKJLzC7Zi6"><p class="footer-contacts-title">Monselice</p><p class="footer-contacts-text">Via Roma 13, Monselice 35043<br/> Padova - Italia<br/> p.iva 05499270287</p></a></div><div class="footer-bottom-right"><ul class="footer-social"><li><a class="mouse-cursor-simple" target="_blank" rel="noopener noreferrer" aria-label="visita il nostro profilo instagram" href="https://www.instagram.com/opificiolamantinianonimi/">Instagram</a></li><li><a class="mouse-cursor-simple" target="_blank" rel="noopener noreferrer" aria-label="visita il nostro profilo linkedin" href="https://it.linkedin.com/company/opificio-lamantini-anonimi">Linkedin</a></li><li><a class="mouse-cursor-simple" target="_blank" rel="noopener noreferrer" aria-label="visita il nostro profilo facebook" href="https://www.facebook.com/OpificioLamantiniAnonimi/">Facebook</a></li><li><a class="mouse-cursor-simple" target="_blank" rel="noopener noreferrer" aria-label="scrivici su whatsapp" href="https://wa.me/39042972134">Whatsapp</a></li></ul><div class="footer-privacy-container"><p class="footer-privacy">Opificio Lamantini Anonimi © <!-- -->2026</p><br/>
65<a class="iubenda-nostyle no-brand iubenda-embed footer-privacy mouse-cursor-simple" title="Privacy Policy" aria-label="leggi la nostra polizza privacy" target="_blank" rel="noopener noreferrer" href="https://www.iubenda.com/privacy-policy/61956652">Privacy Policy</a><a class="iubenda-nostyle no-brand iubenda-embed footer-privacy mouse-cursor-simple" title="Cookie Policy" aria-label="leggi la nostra polizza cookies" target="_blank" rel="noopener noreferrer" href="https://www.iubenda.com/privacy-policy/61956652/cookie-policy">Cookie Policy</a></div></div></div></div></footer></div><div id="cursor" class="" style="left:0px;top:0px"><span class="cursor_txt"></span></div></div>
65<script id="__NEXT_DATA__" type="application/json">{"props":{"pageProps":{"articles":[{"id":632,"title":"Seasonal brands: who knows how to use the emotional calendar (and who fakes it)","subtitle":"Mulino Bianco, Lavazza, and three Italian SMEs that made the calendar a brand asset, not a marketing cost. What they have in common — and why smart seasonality is a positioning move that Italian SMEs underrate.","date":"2026-08-10","quote":"Seasonality isn't a calendar of promotions. It's an emotional grammar. Mulino Bianco has known it for forty years. We forgot, not the customers.","content":"While you're on vacation (or, if you aren't, while half of your team is), your brand is working under the surface. It's connected to the gestures people make in August: opening a fridge in the evening, choosing sunscreen, buying a book for the trip, listening to the radio in the car with the window open.\n\nThese gestures have an **emotional grammar**. There are winter gestures (coffee at mid-morning at the bar), spring gestures (the first outdoor aperitif), autumn gestures (the new notebook for back-to-school), summer gestures (the evening gelato). They are different in emotions, symbols, rhythm.\n\nBrands that knew how to hook into these gestures — not with a seasonal campaign, but with an identity that declines in the seasons — built memory lasting decades. Those that *faked* seasonality (a red pack in December, a graphic with the sun in summer) built boredom.\n\nThis piece is a map of both attitudes. With an operational ending: what you can do, if you're an Italian SME, to take the first path and not the second.\n\n## The two masters: Mulino Bianco and Lavazza\n\n### Mulino Bianco — seasonality as family narrative\n\nMulino Bianco is the most sophisticated Italian case of seasonal brand. Forty years of advertising, packaging, product lines, editorial partnerships — all coherent around one idea: **the seasonality of the Italian family**.\n\nLet's see the details we often miss.\n\nIn September, *the new notebook*. The snack becomes the ritual of going back to school. The packs become more \"binder-like\": more orderly colors, more disciplined illustrations, educational claims.\n\nIn December, *the holidays*. Panettone lines, focus on gift products, coherent red-gold palettes. No shouting. It's festive order, not the party.\n\nAt Easter, *spring in domestic key*. Packaging evokes the home garden, illustrations with animals, themed products (Doria chocolate eggs, Buondì doves).\n\nIn June-July, *the pause*. A communication deceleration: less TV advertising, fewer product launches, some summer line (lighter crackers, fruit cookies). The brand is *still but present*.\n\nIn August, *almost silence*. Very few campaigns, nothing new, the company is there but doesn't insist.\n\nIn September, *we restart*. The cycle begins again, with an amplification coherent with \"back to school\".\n\nFour decades of this coherence built an identity that's practically an *Italian family calendar*. The brand doesn't *follow* seasonality: it *interprets* it. And every gesture of the Italian consumer in the kitchen has a possible point of contact with Mulino Bianco.\n\n### Lavazza — the seasonality of sociality\n\nLavazza makes a different but equally refined move. They don't lean on *weather* seasons, but on the *seasons of sociality*.\n\nLavazza Calendar (the annual photographic calendar, one of the longest-lived institutions of Italian marketing) — different every year, but always tied to an idea of social life, ritual, time flowing gracefully. It's a brand that declares: *coffee is what puts the day on pause, not energy to speed it up.*\n\nSeasonality in campaigns: in September, *back to office* (coffee of the work break). In December, *back home* (the moka on the table, families reunited). In May-June, *out at the bar* (the cappuccino in the piazza). In July, *iced coffee* (summer, deceleration).\n\nUnlike Mulino Bianco, Lavazza doesn't decline on the *domestic gesture* (family lunch), but on the *social gesture* (coffee with another). Two different seasonal grammars, both coherent for decades.\n\n## The three Italian SMEs doing it (almost no one notices)\n\nLet's come to SMEs. Three Italian cases that don't have the budgets of Mulino Bianco or Lavazza, but who built a coherent brand seasonality.\n\n### 1. Vergani Milano — the seasonality of artisanal panettone\n\nVergani is a historic Milanese producer of artisanal panettone. The product is essentially \"Christmasy\" — but they built an annual grammar that avoids the trap of \"disappearing from January to September\".\n\nHow? By distributing the calendar over other themed seasonal products: Easter dove, summer sweet focaccias (Venetian focaccia, Carnival tortelli), autumn products (raisin panettone, holiday focaccia). The brand is always present, always with products the season *wants*. In December it amplifies, but never disappears.\n\n**Lesson:** if your core is seasonal, build *around* the core a rotation of minor products that keep you present all year.\n\n### 2. Pomì — the tomato as summer flag\n\nPomì in recent years has refocused its communication on summer — the natural moment of the tomato. Packs that get more yellow/red in May-June, summer recipes featured, partnerships with festivals and village fairs.\n\nSeptember-October, the brand returns to \"comfort\" mode: tomato as ingredient of cold-season recipes (long sauces, ragù, parmigiana). Different packs, different palette.\n\nWinter: more sober presence, focus on base product. No shouts, no forcings.\n\n**Lesson:** a mono-product brand can still have seasonality, *if* it knows how to tell itself in the different ways the product is used in the seasons.\n\n### 3. Pellini Caffè — coffee as work cycle\n\nPellini (Verona coffee, family-owned) has developed in the last five years a seasonality that leans not on weather but on the *Italian work cycle*.\n\nIn September-October, communication like \"coffee to restart, coffee for the first day\". In December-January, \"coffee for domestic holiday breaks\". In June-July, \"coffee on the road, coffee on summer trips\". In August, *a lot of silence*: the feeling is that they're on pause too.\n\nIt's an *empathic* seasonality. The brand moves at the customer's rhythm. It doesn't force them, doesn't precede them, doesn't follow them: it walks together.\n\n**Lesson:** the most powerful seasonality is not that of the four meteorological seasons. It is that of the seasons *of the customer's life*.\n\n## How to do it in your SME (the practical version)\n\nThree steps.\n\n**1. Define your customer's \"emotional calendar\".**\n\nNot the calendar of your promotions. The calendar of *the customer's emotions*. When does your customer buy lightheartedly? When do they buy under stress? When do they gift? When do they replace? When do they renew?\n\nMap six or seven emotional moments in the year. Each evokes a color, a language, a tone.\n\n**2. Decline your brand in each moment.**\n\nDon't change the logo, don't change the system. But *amplify or decelerate*. For example: at Christmas raise the chromatic volume by 20%, in August lower it by 30%. In September widen the tone of voice toward \"let's restart\", in May toward \"let's allow ourselves a pause\".\n\nIt's an orchestration, not a substitution.\n\n**3. Measure coherence, not innovation.**\n\nThe seasonal brand that works is *predictable*. Your loyal customers expect that in December you'll show up a certain way. If you suddenly change register for \"not being repetitive\", you disappoint them. *Intelligent* repetition is the superpower of seasonality.\n\n## The honest final note\n\nThis piece arrives mid-August, when half of you are on vacation and the other half are doing the September editorial plan. It's deliberate.\n\nAn emotional calendar map is done **best in August**, when you have distance from your brand and can look at it as a customer looks at it. Try to jot it down in 90 minutes, on a notebook, while your team works at half speed. It's one of the most underrated exercises of annual planning.\n\nThursday, August 13 we close the summer editorial plan with the AI pillar long-form: *AI in business: 18-month review and compass for 2027.* What really worked in Italian SMEs, what was hidden cost, where it's worth investing the next 12 months.\n\nSee you Thursday — the last before real rest.\n\n---\n\n**Want to understand if your brand has intelligent seasonality or just a calendar of promotions?** We look at it together on [substantial branding projects](/branding-su-misura-strategia-marketing-aziendale). [Let's talk](/contatti) when you come back.","createdAt":"2026-05-18T10:37:28.803Z","updatedAt":"2026-05-18T10:37:28.803Z","publishedAt":"2026-05-18T10:37:28.819Z","locale":"en","documentId":"w5wuwwctm6r4eoybgoth8x79","slug":"seasonal-brands","pillar":null,"cover":null},{"id":630,"title":"#HardWords: Multimodal","subtitle":"When an AI model 'sees' beyond words: images, audio, video. What changes practically in customer service, e-commerce, manufacturing, and why the difference between text-only AI and multimodal AI is huge for many business cases.","date":"2026-08-06","quote":"Text-only AI is like a specialist who speaks only by email. Multimodal AI is a specialist who can also look at the photo you send, hear the recording, read the floor plan. For many real problems, the difference is everything.","content":"Fourth episode of the series. Today we tackle a 
65word that has become central in 2026 AI presentations and which, for how it gets used, sounds much more technical than it actually is: **multimodal**.\n\nLet's start with the word itself. *Modality* in AI language means \"type of data\": text, images, audio, video. A *unimodal* model works with one type. A *multimodal* model works with two or more types together.\n\nThat's it. It's the basic concept.\n\nWhat makes the thing interesting is not the definition. It's **what becomes possible** when an AI can look, listen, read, and understand — all at the same moment.\n\n## The metaphor\n\nThink of a specialist you can consult by email.\n\nIf your specialist can only read your words, you have to *tell them* everything: the problem, the context, the details, what the object looks like, what sound it makes, where the symptom is. That's what a text-only AI does (base ChatGPT, text Claude). It works, but much of the time is spent *describing* instead of *solving*.\n\nIf your specialist can also *see* the photo you send them, *hear* the audio recording, *read* the floor plan you forward — then a big part of describing disappears. You say *\"this thing makes this noise when I turn it on\"* and send a 5-second video. The specialist knows exactly what you're talking about.\n\nThis is a multimodal AI. They understand you with less effort on your part. And often they give you more precise answers, because they have more context to start from.\n\n## The four business cases where everything changes\n\nThree years ago, these applications were advanced research or demos. Today, in 2026, they are real products you can buy or build.\n\n### 1. Customer service: the user sends the photo\n\nReal example. A customer of an appliance company writes to customer care: *\"the washing machine is leaking water, what do I do?\"*. Classic customer service: ten back-and-forth messages to figure out where it leaks from, how much, in what conditions, model, year of purchase.\n\nMultimodal customer service: the customer sends the photo of the washing machine. The AI recognizes the model, visually identifies the leak (under the gasket, from the filter, from the drain hose), and in two exchanges provides the right procedure.\n\nWhat's needed for it to work: a multimodal model (today easily accessible via API), a knowledge base of the product (see [the RAG hardwords piece](/blog/hardwords-rag)), an integration with the ticketing system.\n\nTypical ROI: average ticket resolution time drops 40-60%. For high-volume companies, it's huge.\n\n### 2. E-commerce: image search\n\nExample. A user sees a sofa in a photo on Instagram, would like to buy it but doesn't know brand/model. Opens your site, uploads the photo. The multimodal AI recognizes visual characteristics — shape, color, structure, fabric, legs — and finds in your products the 5 most similar.\n\nIt works very well for furniture, fashion, accessories. It works less well for categories where the difference is functional and n
65ot visual (one mouse from another mouse).\n\nTypical ROI: conversion rate +20-35% on products where image search is activated. Higher on marketplaces, lower on single-product brands.\n\n### 3. Manufacturing: visual quality control\n\nClassic Italian example. A production line that assembles components. Rare defects, but expensive. Quality control traditionally done by expert operators looking \"by eye\".\n\nMultimodal AI: a camera over the line, a model trained to recognize specific defects. The model flags every suspicious piece to a human operator who decides. The model doesn't replace the operator — *it amplifies them*.\n\nWhat was *random quality control* becomes *systematic quality control*. Defects that previously escaped get caught. False positives get educated.\n\nTypical ROI: depends heavily on the sector, but for Italian precision engineering manufacturing companies — think your [Loccioni](/blog/bravehistories-loccioni-silent-brand-marche) — it's one of the most mature and best-quantifiable cases.\n\n### 4. Compliance and documentation: \"reading complex documents\"\n\nExample for professional studios, consultancies, healthcare. An accountant has to check a balance sheet: 80-page PDF with tables, charts, marginal notes. The text-only model struggles with complex PDFs (tables \"break\" in the text). The multimodal model looks at *the image* of the page and understands the table the way a human would.\n\nSame goes for medical records, contracts with highlighted clauses, floor plans with annotations, technical sheets.\n\nTypical ROI: time spent \"deciphering\" a complex document drops 70-80%. The part where AI can err (a wrong interpretation) is managed with a final human review, which remains indispensable.\n\n## How to know if you need it\n\nA quick checklist. If you answer yes to one of these, multimodal AI is a serious lead for you.\n\n- Do your customers send you photos, videos, audio in daily communications?\n- Do your products have a visually recognizable physical component?\n- Does your work include \"reading\" documents with many images, tables, schemas?\n- Do you have specialized operators doing visual or auditory checks?\n\nIf the answer to all is no, text AI (more mature, cheaper) probably suffices. If the answer to one is yes, a multimodal pilot is worth evaluating.\n\n## When you don't need it\n\nThree red flags.\n\n**They propose \"multimodal AI\" for problems that are text-only.** Resist. Adding modalities you don't need is just complexity and cost.\n\n**They propose a \"custom\" multimodal model without having done a pilot.** Commercial models (GPT-5, Claude 4, Gemini 3) are natively multimodal and work very well on almost all common use cases. Custom is needed only for *really* specific cases (specialized medicine, precision industrial vision). For the rest, start with the commercial model.\n\n**They propose a system that does \"everything\" — text, images, audio, video, advanced reasoning.** It's probably a package sale, not a targeted solution. Start from a use case, scale when the first works.\n\n## Summary in three lines\n\nA multimodal AI is a model that understands more types of data together — text, images, audio, video.\n\nIt changes everything in business cases where communication with the customer or the process isn't only textual.\n\nIt's worth starting from a pilot on a specific case, not from an \"enterprise multimodal AI strategy\".\n\nMonday, August 10 we return to the branding pillar with a piece on **seasonal brands**: who knows how to use the emotional calendar (Mulino Bianco, Lavazza) and who fakes it. Three Italian SMEs that made the calendar an asset, not a cost.\n\n---\n\n**Are you wondering if multimodal AI can really help you?** Under [artificial intelligence for business](/intelligenza-artificiale-per-il-business) we do half-day AI use-case audits. We tell you if it makes sense, and how much it costs to do well. [Let's talk](/contatti).","createdAt":"2026-05-18T10:37:28.387Z","updatedAt":"2026-05-18T10:37:28.387Z","publishedAt":"2026-05-18T10:37:28.408Z","locale":"en","documentId":"ksqjfvnypsntn7uoam7ocmyb","slug":"hardwords-multimodal","pillar":null,"cover":null},{"id":628,"title":"The September pitch: 3 things to fix in August if you have a round in sight","subtitle":"Deck, narrative, numbers. What will change in the next six months among Italian investors — and what they'll ask you first if you open the conversation in September. A preventive checklist to do calmly now that there's no pressure yet.","date":"2026-08-03","quote":"A good round starts three months before you think to start it. Everything you do in August weighs more than the first investor meeting in September.","content":"If you're planning to open a round in Q4 2026 — Series A, Seed, bridge, whatever it is — the useful preparation moment is *now*. Not in September, when you're already sending the first emails. Now, while the company runs at 60%, you have a little headspace, and you can calmly review what you produced hastily during the year.\n\nThree specific things. All doable in two or three days of work distributed in the month. None requires consultants. All make the difference between an investor meeting that proceeds and one that runs aground.\n\n## 1. The deck: cut 30%\n\nThe decks we see passing through in 2026 all have the same problem. **They're too long and say too much.**\n\nThirty-eight slides, each with its title, paragraph, chart. The investor flips through in 4 minutes — it's the declared average time they dedicate to an inbou
65nd deck — and forgets everything the next day.\n\nThe 2026 deck that works has between 12 and 16 slides. One idea per slide. One memorable sentence per slide. Everything else is in the appendix, or told verbally, or not told at all.\n\nExercise for August: take your current deck, count the slides, and put a black cross on the 10-15 slides you can remove without changing the substance. You'll probably notice you can remove more than you thought. When you've finished, look at it again. If it's still too long, cut more.\n\nA heuristic rule: if the slide doesn't add an *information that changes the reader's opinion*, it must come out. *\"We also do X\"* doesn't change opinion. *\"We're the only ones in Italy doing X this specific way\"* does.\n\n## 2. The narrative: the story in 90 seconds\n\nRegardless of the deck, you must have a **90-second oral narrative** of your business. Ready. By heart. Natural.\n\nNot a one-minute pitch memorized *\"we're the Uber of X\"* style. A real story. Beginning, development, where you are today, where you're going.\n\nExample structure:\n\n- **Beginning**: the problem that made you meet your market. *\"Three years ago we were looking for a solution to X, we didn't find it, and we understood the problem was more widespread than we thought.\"*\n- **Development**: what you built and why in that specific way. *\"We tried approach Y for six months, understood it didn't scale, and went back to logic Z — which today is our competitive advantage.\"*\n- **Today**: the present situation, in concrete numbers. *\"Forty paying customers, ARR of 800k, growth of 15% month-on-month on enterprise customers.\"*\n- **Where you're going**: the promise for the next 18-24 months. *\"We want to reach 3 million ARR by bringing the model to another three adjacent sectors. That's what we're opening the round for.\"*\n\n90 seconds. Spoken. Without notes. From when you enter the investor's room, you must be able to say this thing naturally — it's the *first* test the investor gives you, even before looking at the deck.\n\nPractice it. Literally. With a stopwatch. Record yourself. Listen back. It's boring, but it's where the first 30% of the valuation is played.\n\n## 3. The numbers: three tables, not twenty charts\n\nItalian investors — and the European ones who flank them — in 2026 ask for *fewer* metrics than a year ago, but *much more seriously*. They've moved from the \"wow factor\" phase to the \"due diligence anticipated at the first meeting\" phase.\n\nThe three tables that must be **ready, updated to the previous month, and immediately defensible** are:\n\n**Table 1 — Complete commercial funnel.** From qualified lead to signed customer to active customer, month by month, last 18 months. Absolute numbers, not percentages (they calculate themselves). If you don't have 18 months of data, put what you have and declare it honestly.\n\n**Table 2 — Unit economics.** Customer acquisition cost (CAC), customer lifetime value (LTV), payback period, gross margin. For each main customer segment. Even if the numbers are bad (high CAC, still low LTV because you're young), put them — *a serious investor prefers honest bad numbers to clean fake numbers*.\n\n**Table 3 — Cash runway.** How many months do you have at the current spending pace? What changes if growth slows? What changes if you increase sales? Three scenarios (base, prudent, aggressive), all on one page.\n\nThese three tables go *outside* the deck, in an appendix or separate Excel. They serve in follow-up conversations, but they go prepared before the first meeting. If the investor asks for them and you reply \"we have them, I'll send by Wednesday\" — you lose speed. If you send them within 30 minutes of the call — you gain credibility.\n\n## Four things NOT to do in August\n\nFor honesty, also what to avoid.\n\n**Don't send emails to investors in August.** 90% won't reply until mid-September. The 10% who reply will be distracted. You waste first-impression chances.\n\n**Don't completely rewrite the deck from scratch.** Iterate on what you have. Revision pays more than restart.\n\n**Don't try to \"perfect traction\".** Traction is made by product and team, not by August. If the numbers aren't sufficient now, they won't be on September 15 due to a month of extra work.\n\n**Don't talk about your round with everyone.** Voices cir
65culate. Having \"round opening\" as market rumor without yet a serious plan is a disadvantage. Keep everything in the family (founder + max 2 advisors) until September.\n\n## A note on how we're seeing valuations change\n\nWhat we observe in rounds closed in the first half of 2026: Italian Series A valuations are **about 30% below the 2021-2022 peaks** but they're *stabilized*. The downward pressure of 2024 is no longer seen.\n\nWhat does it mean for you: if you open a round in September, start with realistic expectations. Valuation multiples on ARR have normalized in the 6-10x range for B2B SaaS (down from 15-20x in 2021). For applied AI it goes higher, but with very strict conditions on milestones. For consumer / DTC it struggles.\n\nIndicative numbers, varies a lot by sector and geography. But the strategic message is: **evaluate what valuation you've mentally prepared for, and drop by 25-30%**. That will be the realistic starting base of the conversation.\n\nThursday, August 6 we return to the glossary with a much-requested episode: **#HardWords: Multimodal.** When a model \"sees\" beyond words — and what changes practically in customer service, e-commerce, and manufacturing.\n\n---\n\n**Are you preparing a round for Q4?** Under [brand and narrative strategies for startups](/startup-innovative-sviluppo-su-misura), we help founders build the story they sell to investors. [Let's talk](/contatti) — even in August, even at the end of August.","createdAt":"2026-05-18T10:37:28.036Z","updatedAt":"2026-05-18T10:37:28.036Z","publishedAt":"2026-05-18T10:37:28.054Z","locale":"en","documentId":"xbb2fu0je9t5n8z86qwnfd9v","slug":"september-pitch-3-things-to-fix","pillar":null,"cover":null},{"id":626,"title":"The end of 'easy' SEO traffic: how to rebuild a realistic-driven plan","subtitle":"Keyword volumes have become unreliable indicators. Zero-click searches have eaten easy traffic. A methodology to redesign the content plan on intents, opportunity cost, and a new way to measure impact.","date":"2026-07-30","quote":"For fifteen years we treated keyword volume as Gospel. In 2026 it's little more than a historical statistic with a panoramic view. Continuing to use it as the only criterion is like deciding what to eat by looking at a three-year-old restaurant menu.","content":"For fifteen years, an average company's content plan was written more or less like this. You open Ahrefs (or SEMrush, or Ubersuggest). You search for keywords with good volume and low difficulty. You write a list of \"SEO opportunities\". You assign them to freelance copywriters or to the agency. You publish.\n\nIt worked. More or less. For the right sectors, even well.\n\nIn 2026 this approach is dead, and almost no one has announced it.\n\nThe problem isn't that keywords don't exist anymore. It's that **their volume no longer corresponds to the traffic you get by ranking on them**. Between zero-click searches (today over 60% of the total), generative AI answers, expanded snippets, sponsored cards, and local maps, the first organic result of a query with \"monthly volume 5,000\" today brings — at best — a few hundred visits.\n\nResult: content plans written on Ahrefs continue to look valid. They produce well-ranked articles. And they don't bring traffic.\n\nThis piece is a reconstruction methodology. No theory. Five practical steps to redesign your content plan for the next quarter on a different criterion: **intent and opportunity cost**, no longer volume.\n\n## Step 1: clean up the metrics of old content\n\nBefore planning what to write, understand what you've already written and how it really performs.\n\nOpen Search Console. Filter the articles from your blog from the last 24 months. Export the dataset. For each article, calculate:\n\n- **Real clicks** (not impressions, real clicks)\n- **CTR** (clicks / impressions)\n- **Average position**\n- **Article age**\n\nSort by descending clicks. You'll probably notice two things: 20% of articles bring 80% of traffic (it's almost always like this), and among that 20% many are *old* (3-5 years).\n\nThat pattern is precious information. It means:\n\n- The content of yours that works is little\n- It's almost always the **deep** content (long-form, evergreen, well-cared-for over time)\n- The \"shallow\" content written for a keyword opportunity is producing little or nothing\n\nFrom this alone, a decision can be made immediately: **stop writing \"shallow\" articles** (800 words, generic listicle, written for a keyword). Invest that time in two things: updating those that work, and writing few new ones but *very* deeply.\n\n## Step 2: identify intents, not keywords\n\nKeyword research in 2026 works like this. Instead of starting from a keyword and asking *how many search for it*, start from a **customer intent** and ask *how do they search for it*.\n\nExamples. If you sell industrial machines, your customer's intent isn't \"search for the keyword *machine X*\". It is:\n\n- *\"Understand whether to buy machine X or rent it\"*\n- *\"Evaluate if machine X fits their production\"*\n- *\"Compare machine X with model Y\"*\n- *\"Find a machine X installer in Italy\"*\n- *\"Solve a specific problem with machine X\"*\n\nFive distinct intents. Each corresponds to *many* possible queries — some with high volume, some with low volume, some that don't even appear in standard tools. But all five correspond to a **decision moment** of the customer.\n\nThe right question is no longer \"how much traffic does keyword X bring me?\". It's \"
65if my customer *lives this intent*, do I intercept it?\".\n\nFor each intent, write one and only one really good article. Five intents = five articles. Five well-written articles per year beat fifty articles written for keywords.\n\n## Step 3: build the \"opportunity cost\" of each topic\n\nHere comes the counterintuitive part.\n\nFor each candidate topic, calculate **two things** most traditional SEOs ignore:\n\n**A — Real writing difficulty.** How much does it cost to write it *well*? A banal article gets written by the agency in three hours. A truly deep article on a topic you know deeply takes a day of thinking + half a day of writing. If you don't have that time, it doesn't pay you to write it.\n\n**B — Distance from your core value.** Is the topic central to your brand, or peripheral? A well-written peripheral article brings traffic but doesn't position you. A well-written central article brings less traffic but positions you — and in the next years, in a SEO + AEO + GEO world ([see the June 29 piece](/blog/from-seo-to-geo)), positioning matters more than traffic.\n\n**Opportunity cost** = (potential volume) × (closeness to core) ÷ (writing difficulty).\n\nIt's an approximate formula, but it changes priorities. Often you'll notice that topics with higher volumes have low opportunity costs (because they're peripheral and difficult). And topics with mediocre volume have high opportunity costs (because they're central and relatively simple to write well).\n\nPlan on the second category. Not the first.\n\n## Step 4: stop measuring only traffic\n\nKPI change. For each new article, define the **expected result** in three dimensions, not one:\n\n1. **Clicks** (the traditional metric).\n2. **Citation in AI chatbots** (does the article get included when an LLM answers questions on the topic? Verifiable manually with ChatGPT, Perplexity, Claude.)\n3. **Qualified conversion** (those who read the article sign up, subscribe, request a demo, at what percentage?).\n\nThe \"ideal\" article of 2026 is the one that wins on at least *two out of three* axes. Articles that win only on clicks but not on citation and not on conversion are *vanity content* — they look effective but don't move the business.\n\nInitially it's laborious. Three months later, you have a dashboard telling you which content is *really* building positioning for you.\n\n## Step 5: use AI as *editor*, not as *author*\n\nA final operational note. In 2026 vibe writing — using AI to write blog articles — is everywhere. Everyone does it. Result: the public web is invaded by AI-written articles, and Google's algorithms + LLMs' citation models are *learning to re
65cognize and devalue them*.\n\nThe operational rule we apply internally and recommend:\n\n- AI is excellent for **structuring** an article (outline, angle definition, point map).\n- AI is excellent for **reviewing** an already-written article (flow check, redundancies, cut suggestions).\n- AI is excellent for **expanding research** (finding sources, lining up data, generating examples).\n- AI **doesn't replace** the author's voice. Voice — point of view, opinion, \"why I say it\" — must be human.\n\nAn article where 100% of text is AI-generated is visible three steps away today. An article where AI did 30-50% of the support work (research, outline, review) and the human put voice, opinion, and angle is the winning formula of 2026-2027.\n\n## The final truth\n\nCompanies seriously doing SEO in 2026 changed the starting point of the content plan. No longer \"what do our customers search for?\". But rather: **\"what do we have to say that no one else knows how to say as we do?\"**.\n\nIt sounds silly. It is instead the difference between chasing Google's 2015 algorithm (which rewards those who answer generic questions) and speaking to the 2026 reader-customer (who wants informed opinions, data, direct experience).\n\nItalian SMEs have a huge advantage on this axis: they have **real experiences** that \"optimized\" articles never have. Putting those experiences on the page is the best SEO you can do today.\n\nMonday, August 3 we enter real August with the first of the pre-Q4 pieces: *The September pitch: 3 things to fix in August if you have a round in sight.* Deck, narrative, numbers — what will change in the next six months among Italian investors.\n\n---\n\n**Have an editorial plan that produces articles but not results?** Under [growth hacking and sustainable organic growth](/growth-hacking-strategie-per-crescita-sostenibile) we rebuild content strategies based on intents, not keywords. [Let's talk](/contatti) — September is the right month to restart.","createdAt":"2026-05-18T10:37:27.685Z","updatedAt":"2026-05-18T10:37:27.685Z","publishedAt":"2026-05-18T10:37:27.696Z","locale":"en","documentId":"wfidoxbr82e0prhpmb6yr9v3","slug":"the-end-of-easy-seo-traffic","pillar":null,"cover":null}],"article":{"id":634,"title":"AI in business: 18-month review and compass for 2027","subtitle":"Eighteen months of generative AI applied to Italian SMEs. What worked, what was hidden cost, where to invest in the next 12 months. The long-form closing of the summer — to re-read at Christmas to see if we got it right.","date":"2026-08-13","quote":"2024 was the year of the promise. 2025 was the year of disappointment. 2026 is the year of discipline. Those who understand this trajectory survive the next 18 months without ruining themselves.","content":"In early 2025, a boardroom of an Italian SME in the Northeast. Four people: CEO, CFO, IT head, marketing. The agenda item is \"define the company's AI strategy\". On the table is a stack of printed newsletters, a consulting firm's 80-page report, two quotes from different vendors, and the weekend edition of Sole 24 Ore with a full-page headline on generative AI.\n\nThe meeting lasts three hours. They come out with a decision: \"evaluate over the next six months\". In June 2025 a plan is presented. In September 2025 it's postponed. In January 2026 it's resurrected. In May 2026 a budget is approved. In September 2026 — now — implementation will begin.\n\nSeventeen months.\n\nYes, the company is a real case. No, we can't name it. But it's exactly *the average Italian company* — among the SMEs that did something on AI in 2024-2026.\n\nThis piece is a long-form review. Eighteen months of generative AI applied to Italian companies, distilled into what we've seen work, what turned out to be hidden cost, and where it's worth investing in the next 12 months. It's deliberately long. It's deliberately frank. It's deliberately written **to be re-readable at Christmas** — because many things said here might already be outdated, and those that won't be will be our shared compass for 2027.\n\n## Three phases, one trajectory\n\nIn 18 months, AI in business went through three clear phases.\n\n**Phase 1 — The promise (January 2024 — June 2025).** Everyone wants to do AI. Nobody knows exactly what to do with it. Budgets inflate, quotes proliferate, \"ChatGPT for managers\" courses multiply. AI strategy is a statement of intent: *\"we want to be AI-first\"*. Underneath, little concrete.\n\n**Phase 2 — The disappointment (July 2025 — December 2025).** First pilots go badly or go *less well than expected*. Companies discover that integrating an LLM into their operations is more complex than the demo suggested. CMOs who pushed AI projects come under pressure from the CFO. Some initiatives get frozen, others reduced. *\"AI deception\"* becomes common talk.\n\n**Phase 3 — Discipline (January 2026 — today).** Companies that survived phase 2 stopped asking \"should we do AI?\" and started asking \"*for which specific problem is AI the best solution?*\". Pilots become smaller, more targeted, more measurable. The vocabulary refines: *generative AI*, *agentic AI*, *RAG*, *context engineering*, *multimodal* are distinguished. Decisions are made with discipline.\n\nThe three phases are not over. They coexist. Every new company entering the AI market lives the three phases *in compressed scale*: from promise to discipline in a few months if lucky, in two years if less attentive. But the trajectory is clear, and it's that one.\n\n## What really worked (in Italian SMEs)\n\nWe give you the real cases, anonymized but recognizable as patterns. They are the AI applications where we saw **verified ROI** in Italian SMEs.\n\n**1. AI-assisted (not replaced) customer service.** For companies with recurring weekly ticket volumes (services, e-commerce, software), an AI assistant that *prepares* human responses (doesn't send them directly) reduces average handling time by 40-60%. It works because it preserves human responsibility, but eliminates repetitive typing. Investment: medium. Time to return: 3-6 months.\n\n**2. Internal data research and analysis.** Companies with lots of internal documentation (legal, consulting, manufacturers with technical manuals) found great value in RAG systems (see [the hardword piece on RAG](/blog/hardwords-rag)) that allow querying the archive in natural language. It works because it solves a real problem — *\"I have a specific question and the answer is in one of our 4,000 documents\"*. Investment: medium. Time to return: 6-9 months.\n\n**3. Creative variant generation.** Marketing teams managing many campaigns used AI to generate copy and creative ad variants, accelerating testing. Doesn't replace the creative. *Amplifies them.* Works very well for ambient marketing, performance marketing, social ads. Investment: low. Time to return: 1-3 months.\n\n**4. Accelerated software development (with rules).** As written in the [July 23 vibe coding piece](/blog/vibe-coding-in-production), internal development teams got +50-80% release speed. *With governance.* Investment: medium-high (in training and processes). Time to return: 6 months.\n\n**5. Complex document audit.** Professional firms and consultancies are using multimodal AI for first passes on balance sheets, contracts, technical sheets. Human review remains indispensable, but initial \"decoding\" time collapses. Works very well for accounting firms, mid-sized law firms, tax consultants. Investment: low. Time to return: 3 months.\n\n## What was hidden cost\n\nNow the part less 
65willingly told. Three things that, in many SMEs, turned out to be expensive without anyone noticing first.\n\n**1. Time spent \"evaluating vendors\".** Italian companies above a certain size spent dozens of half-days of the CFO, IT head, marketing, in meetings with AI vendors in 2024-2025. The vast majority of these meetings produced no decisions. Submerged opportunity cost: easily 30-50 thousand euros of managerial time per average company. On a thousand Italian SMEs, we're talking about **30-50 million euros a year** of managerial time burned on evaluations. Few notice because it's not in any cost center.\n\n**2. Overlapping \"tool stacks\".** Many companies today pay simultaneously: ChatGPT Enterprise licenses, Claude Team licenses, Copilot licenses for Microsoft 365, and one or two vertical tools (for chatbots, for analysis, for coding). Aggregate cost: often 30-60 thousand euros a year for a 50-employee company. Effective usage: ridiculously low. Rationalizing the tool stack is one of the most profitable exercises of Q4 2026.\n\n**3. Late-managed compliance.** As written in [the June 25 piece on AI Act phase 2](/blog/ai-act-phase-2), many companies discovered *after* putting AI systems into production that they had to document, evaluate, and supervise. The cost of doing it *after* is triple compared to doing it *before*. Italian SMEs that structured AI governance practices already in 2025 are saving money and time in 2026.\n\n## The compass for 2027\n\nLet's come to the prediction part. Four trends we see consolidating, and which you should budget for in the next 12 months.\n\n### Trend 1 — AI becomes \"vertical\", not \"horizontal\"\n\nFor the first two years of generative AI in business, the dominant tools were horizontal: ChatGPT, Claude, Gemini, Copilot. Tools that do *many things mediocrely well*.\n\nIn 2027 value shifts to *vertical* tools. AI systems built specifically for: accounting, legal management, manufacturing, healthcare, consulting, education. They work *much better* for the specific case — because they're designed around it — and in return *they only work for that*.\n\n**What to do:** evaluate what you really need, and look for the *vertical* vendor that solves it. Stop thinking \"ChatGPT is the universal base\".\n\n### Trend 2 — AI agents start replacing workflows (not people)\n\nAgentic AI — systems that execute chains of actions autonomously — is moving from promise to operability in the next 12 months. But the way it replaces people is different from what everyone thought.\n\nIt doesn't replace customer service. It replaces *the customer service workflow* — the three different tools the operator uses to handle a case. The operator remains, but their work environment simplifies radically.\n\n**What to do:** identify the most *friction-filled* workflows in your company (those requiring passages between many tools, many hands, many waits). Agents will enter there. Not elsewhere.\n\n### Trend 3 — \"AI-native\" become a new segment\n\nIn 2027 we'll see companies emerge that are **AI-native from birth**: organizations built with the assumption that AI is a constitutive part of the process, not an addition. It's what we're trying to do with [OpificioAI](https://opificio.ai), Tuken, BP, BI, and Blogger. We don't do it for fashion — we do it because we understood, in these 18 months, that AI-native companies *operate in a structurally different way*. They have smaller teams, faster decisions, products that scale with almost zero marginal cost.\n\n**What to do:** if you're planning a new business unit, evaluate building it AI-native from the start. On existing products it's difficult to retrofit. On new ones it's the right move.\n\n### Trend 4 — Trust becomes the real asset\n\nIn 2026 we saw a strange phenomenon: some technically *worse* AI vendors won contracts against technically *better* vendors. Why? Because they knew how to build *trust*. Data transparency, cost clarity, ready compliance documentation, open audit processes.\n\nIn 2027 this trend will amplify. Companies will increasingly buy AI from vendors who can **demonstrate**, not just **declare**, to be reliable. Trust becomes part of the product.\n\n**What to do:** if you're an AI vendor, invest in transparency and documentation now. If you're a customer, choose vendors based on transparency, not just performance.\n\n## Five operational tips for the next quarter\n\nTo close, five concrete actions for Q4 2026. All doable in 30-90 days.\n\n1. **AI tool stack audit.** How many tools do you pay for? How used are they? What can you consolidate? Typically: -30% annual costs, +20% clarity.\
65n\n2. **Audit of AI systems in production.** How many are in \"high risk\" tier per the AI Act? For each, is there a compliance plan? Involve legal. It's done in half a day, avoids millions of problems.\n\n3. **Define the product metric, not the agent.** For each AI system in the company, what's the *business* metric telling you if it's working? If you don't have it, build it. A clear metric beats a thousand engineering dashboards.\n\n4. **Invest in governance before the next tool.** House rules (see [the May 28 piece on the 4 recurring mistakes](/blog/ai-agents-in-production-4-recurring-mistakes)) are worth more than a new model, a new agent, a new license. Spend there.\n\n5. **Start a single vertical pilot, not a horizontal one.** See trend 1. Identify the most specific problem in your company where AI can help, and put three months on it. More focus, more ROI, more learning.\n\n## The final note for those reading us\n\nThis is the last piece of the May-August 2026 editorial plan. Twenty-six articles, written two a week, without skipping any. A promise we made you in May when we reopened the blog after seven months of silence, and that we kept.\n\nIn September we'll be back with a new editorial block. Topics, cadence, tone — we'll define them in the coming weeks. If you have suggestions, write to us. This blog isn't a monologue, it's a conversation.\n\nHappy Ferragosto. Good return. See you in September, fresh, with some more ideas than now. And maybe — with a little discipline and a good dose of curiosity — one more piece of AI that really works in our SMEs.\n\n---\n\n**Do you feel halfway between the disappointment phase and the discipline phase?** Under [artificial intelligence for business](/intelligenza-artificiale-per-il-business) and with [OpificioAI](https://opificio.ai) we help Italian SMEs choose and put into production the AI that makes a difference. [Let's talk](/contatti) in September, when you're really ready.","createdAt":"2026-05-18T10:38:10.405Z","updatedAt":"2026-05-18T10:38:10.405Z","publishedAt":"2026-05-18T10:38:10.425Z","locale":"en","documentId":"s4ptgcya6xtpr1zvwxcl0ico","slug":"ai-in-business-18-month-review-2027-compass","pillar":null,"cover":null,"audio":null},"locale":"en"},"__N_SSG":true},"page":"/blog/[articolo]","query":{"articolo":"ai-in-business-18-month-review-2027-compass"},"buildId":"Wk-8BYbl35XFN6ORArdsv","isFallback":false,"isExperimentalCompile":false,"gsp":true,"locale":"en","locales":["it","en"],"defaultLocale":"it","scriptLoader":[{"id":"segment","strategy":"afterInteractive","children":"\n                  !function(){var i=\"analytics\",analytics=window[i]=window[i]||[];if(!analytics.initialize)if(analytics.invoked)window.console\u0026\u0026console.error\u0026\u0026console.error(\"Segment snippet included twice.\");else{analytics.invoked=!0;analytics.methods=[\"trackSubmit\",\"trackClick\",\"trackLink\",\"trackForm\",\"pageview\",\"identify\",\"reset\",\"group\",\"track\",\"ready\",\"alias\",\"debug\",\"page\",\"screen\",\"once\",\"off\",\"on\",\"addSourceMiddleware\",\"addIntegrationMiddleware\",\"setAnonymousId\",\"addDestinationMiddleware\",\"register\"];analytics.factory=function(e){return function(){if(window[i].initialized)return window[i][e].apply(window[i],arguments);var n=Array.prototype.slice.call(arguments);if([\"track\",\"screen\",\"alias\",\"group\",\"page\",\"identify\"].indexOf(e)\u003e-1){var c=document.querySelector(\"link[rel='canonical']\");n.push({__t:\"bpc\",c:c\u0026\u0026c.getAttribute(\"href\")||void 0,p:location.pathname,u:location.href,s:location.search,t:document.title,r:document.referrer})}n.unshift(e);analytics.push(n);return analytics}};for(var n=0;n\u003canalytics.methods.length;n++){var key=analytics.methods[n];analytics[key]=analytics.factory(key)}analytics.load=function(key,n){var t=document.createElement(\"script\");t.type=\"text/javascript\";t.async=!0;t.setAttribute(\"data-global-segment-analytics-key\",i);t.src=\"https://cdn.segment.com/analytics.js/v1/\" + key + \"/analytics.min.js\";var r=document.getElementsByTagName(\"script\")[0];r.parentNode.insertBefore(t,r);analytics._loadOptions=n};
65analytics._writeKey=\"CSJmcHfP5yffpzDJLdwHBep2hP30zu26\";;analytics.SNIPPET_VERSION=\"5.2.0\";\n                  analytics.load(\"CSJmcHfP5yffpzDJLdwHBep2hP30zu26\");\n                  analytics.page();\n                  }}();\n                "},{"strategy":"afterInteractive","src":"https://www.googletagmanager.com/gtag/js?id=AW-11081991490"},{"id":"google-aw","strategy":"afterInteractive","children":"\n                  window.dataLayer = window.dataLayer || [];\n                  function gtag(){dataLayer.push(arguments);}\n                  gtag('js', new Date());\n\n                  gtag('config', \"AW-11081991490\");\n                "},{"id":"facebook-pixel","strategy":"afterInteractive","children":"\n                  !function(f,b,e,v,n,t,s)\n                  {if(f.fbq)return;n=f.fbq=function(){n.callMethod?\n                  n.callMethod.apply(n,arguments):n.queue.push(arguments)};\n                  if(!f._fbq)f._fbq=n;n.push=n;n.loaded=!0;n.version='2.0';\n                  n.queue=[];t=b.createElement(e);t.async=!0;\n                  t.src=v;s=b.getElementsByTagName(e)[0];\n                  s.parentNode.insertBefore(t,s)}(window, document,'script',\n                  'https://connect.facebook.net/en_US/fbevents.js');\n                  fbq('init', '655181521892520');\n                  fbq('track', 'PageView');\n                "},{"id":"linkedin","strategy":"afterInteractive","children":"\n                  _linkedin_partner_id = \"1787433\";\n                  window._linkedin_data_partner_ids = window._linkedin_data_partner_ids || [];\n                  window._linkedin_data_partner_ids.push(_linkedin_partner_id);\n                "},{"id":"linkedin2","strategy":"afterInteractive","children":"\n                  (function(){var s = document.getElementsByTagName(\"script\")[0];\n                  var b = document.createElement(\"script\");\n                  b.type = \"text/javascript\";b.async = true;\n                  b.src = \"https://snap.licdn.com/li.lms-analytics/insight.min.js\";\n                  s.parentNode.insertBefore(b, s);})();\n                "},{"id":"iubenda","strategy":"lazyOnload","children":"\n                  var _iub = _iub || [];\n                  _iub.csConfiguration = {\"perPurposeConsent\":true,\"siteId\":1462938,\"whitelabel\":false,\"cookiePolicyId\":61956652,\"lang\":\"it\", \"banner\":{ \"acceptButtonDisplay\":true,\"backgroundColor\":\"white\",\"closeButtonRejects\":true,\"customizeButtonDisplay\":true,\"explicitWithdrawal\":true,\"fontSizeBody\":\"12px\",\"listPurposes\":true,\"position\":\"float-bottom-left\",\"rejectButtonDisplay\":true,\"showPurposesToggles\":true,\"slideDown\":false,\"textColor\":\"black\" }};\n                "},{"type":"text/javascript","src":"//cdn.iubenda.com/cs/ccpa/stub.js","strategy":"afterInteractive"},{"type":"text/javascript","src":"//cdn.iubenda.com/cs/iubenda_cs.js","strategy":"afterInteractive"}]}</script>
65</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.