https://www.dfx5.com/assets/caseStudies-DcB7J9LZ.js

1import{c as a,H as n,C as e,S as i,G as o,B as t}from"./index-4VmgdH68.js";/**
2 * @license lucide-react v0.462.0 - ISC
3 *
4 * This source code is licensed under the ISC license.
5 * See the LICENSE file in the root directory of this source tree.
6 */const s=a("Building2",[["path",{d:"M6 22V4a2 2 0 0 1 2-2h8a2 2 0 0 1 2 2v18Z",key:"1b4qmf"}],["path",{d:"M6 12H4a2 2 0 0 0-2 2v6a2 2 0 0 0 2 2h2",key:"i71pzd"}],["path",{d:"M18 9h2a2 2 0 0 1 2 2v9a2 2 0 0 1-2 2h-2",key:"10jefs"}],["path",{d:"M10 6h4",key:"1itunk"}],["path",{d:"M10 10h4",key:"tcdvrf"}],["path",{d:"M10 14h4",key:"kelpxr"}],["path",{d:"M10 18h4",key:"1ulq68"}]]);/**
7 * @license lucide-react v0.462.0 - ISC
8 *
9 * This source code is licensed under the ISC license.
10 * See the LICENSE file in the root directory of this source tree.
11 */const r=a("Pickaxe",[["path",{d:"M14.531 12.469 6.619 20.38a1 1 0 1 1-3-3l7.912-7.912",key:"we99rg"}],["path",{d:"M15.686 4.314A12.5 12.5 0 0 0 5.461 2.958 1 1 0 0 0 5.58 4.71a22 22 0 0 1 6.318 3.393",key:"1w6hck"}],["path",{d:"M17.7 3.7a1 1 0 0 0-1.4 0l-4.6 4.6a1 1 0 0 0 0 1.4l2.6 2.6a1 1 0 0 0 1.4 0l4.6-4.6a1 1 0 0 0 0-1.4z",key:"15hgfx"}],["path",{d:"M19.686 8.314a12.501 12.501 0 0 1 1.356 10.225 1 1 0 0 1-1.751-.119 22 22 0 0 0-3.393-6.319",key:"452b4h"}]]),c="/assets/humano-architecture-9IeqvlMC.jpg",l="/assets/case-humano-fraud-DgCHdcga.jpg",d="/assets/rappi-delivery-3uzuL8JN.jpg",u="/assets/ansa-mcal-C1F-CIsx.jpg",m="/assets/sdlc-agentic-ai-CeBAH1Qn.jpg",g="/assets/mining-maintenance-agentic-ai-D2VXFB4x.jpg",h=[{slug:"rappi-ai-contact-center",title:"AI-Powered Contact Center for Rappi",icon:n,accent:"from-[hsl(10,80%,55%)] to-[hsl(30,90%,55%)]",heroImage:d,outcome:"Reduced average handle time by 40% and increased first-contact resolution to 92% with AI-driven self-service.",metrics:[{value:"40%",label:"AHT Reduction"},{value:"92%",label:"FCR Rate"}],challenge:"Rappi's contact center handled millions of monthly interactions across multiple countries. Legacy IVR systems created long wait times, high agent attrition, and inconsistent customer experiences across channels.",solution:"DFX5 designed and deployed an Amazon Connect-based omnichannel contact center with AI-driven self-service bots, real-time sentiment analysis, and intelligent routing — reducing agent workload while boosting customer satisfaction.",phases:[{title:"Phase 1 – Discovery & Assessment",bullets:["Mapped existing IVR flows and pain points across 9 countries","Defined KPIs: AHT, FCR, CSAT, and self-service containment","Designed target-state architecture on Amazon Connect"]},{title:"Phase 2 – Build & Integration",bullets:["Built AI self-service bots with Amazon Lex and Bedrock","Integrated with Rappi's CRM and order management systems","Deployed real-time sentiment analysis for agent assist"]},{title:"Phase 3 – Pilot & Optimization",bullets:["Launched pilot in 2 countries with 500 agents","Fine-tuned bot intents and routing rules based on live data","Achieved 60% self-service containment rate in pilot"]},{title:"Phase 4 – Full Rollout",bullets:["Scaled to 9 countries and all support channels","Continuous model retraining with production data","Ongoing managed services and 24/7 monitoring"]}],results:{operational:["40% reduction in average handle time","92% first-contact resolution rate","60% of interactions handled via self-service"],financial:["30% reduction in contact center operating costs","ROI achieved within the first 6 months","Reduced agent attrition by 25%"],customerExperience:["CSAT improved from 3.2 to 4.5 out of 5","Average wait time reduced from 8 minutes to under 1 minute","Omnichannel consistency across chat, voice, and social"]},industry:"On-demand delivery",region:"LATAM (9 countries)",techStack:[{category:"Contact Center",items:["Amazon Connect","Contact Lens","Amazon Lex"]},{category:"AI / ML",items:["Amazon Bedrock","Amazon Comprehend","SageMaker"]},{category:"Data & Integration",items:["Kinesis","Lambda","EventBridge","DynamoDB"]},{category:"Observability",items:["CloudWatch","QuickSight","OpenSearch"]}],pullQuote:{quote:"DFX5 helped us re-imagine customer support as an AI-native operation, not a cost center.",author:"VP of Customer Experience",role:"Rappi"},related:["humano-seguros-fraud-detection","auna-salud-genai-surgical-docs"]},{slug:"cloud-migration-global",title:"Cloud Migration for Global Companies",icon:e,accent:"from-[hsl(210,80%,55%)] to-[hsl(260,70%,55%)]",heroImage:"https://images.unsplash.com/photo-1451187580459-43490279c0fa?w=600&h=400&fit=crop",outcome:"Migrated 500+ workloads to AWS in 6 months, cutting infrastru

11cture costs by 35% while improving uptime to 99.99%.",metrics:[{value:"500+",label:"Workloads Migrated"},{value:"35%",label:"Cost Reduction"}],challenge:"A multinational enterprise operated across fragmented on-premises data centers with aging infrastructure, inconsistent security postures, and escalating maintenance costs that hindered digital transformation.",solution:"DFX5 executed a large-scale cloud migration using AWS Migration Hub, Application Migration Service, and Database Migration Service — modernizing workloads with a lift-shift-optimize approach while maintaining zero-downtime operations.",phases:[{title:"Phase 1 – Assessment & Planning",bullets:["Inventoried 500+ workloads across 4 data centers","Classified applications by migration strategy (6Rs)","Defined migration waves and dependency mapping"]},{title:"Phase 2 – Foundation & Landing Zone",bullets:["Built multi-account AWS landing zone with AWS Control Tower","Implemented network connectivity and security guardrails","Established CI/CD pipelines for infrastru

11cture as code"]},{title:"Phase 3 – Migration Execution",bullets:["Migrated workloads in 12 waves over 6 months","Performed live database migrations with minimal downtime","Validated each wave with automated testing suites"]},{title:"Phase 4 – Optimization & Governance",bullets:["Right-sized instances and implemented savings plans","Deployed centralized monitoring and cost management","Established cloud governance and FinOps practices"]}],results:{operational:["500+ workloads migrated in 6 months","99.99% uptime achieved post-migration","Deployment frequency increased by 300%"],financial:["35% reduction in infrastructure costs","Eliminated $2M in annual data center maintenance","ROI achieved within the first year"],customerExperience:["Application response times improved by 45%","Zero unplanned downtime during migration","Developer productivity increased by 40%"]}},{slug:"humano-seguros-fraud-detection",title:"Fraud Detection for HUMANO Seguros",icon:i,accent:"from-[hsl(142,72%,45%)] to-[hsl(180,60%,45%)]",heroImage:l,outcome:"Deployed real-time fraud detection across healthcare claims, increasing detection by 250% and reducing processing time from 9 hours to 2 minutes.",metrics:[{value:"250%",label:"Detection Increase"},{value:"2 min",label:"Processing Time"}],challenge:"Humano Seguros, a leading health insurance provider in the Dominican Republic with 53% market share and 430K insured members, operated a manual rules-based fraud detection system that took an average of 14 days to identify fraudulent claims and detected less than 70% of fraud — a critical problem given the 25% annual growth in claims volume.",solution:"DFX5 built a multi-stage AI architecture leveraging Amazon Bedrock for document categorization, Amazon Textract for metadata extraction, CNN models on Amazon SageMaker for pixel-level forgery detection, and Amazon OpenSearch for physician credential verification — all orchestrated via AWS Step Functions for scalable, real-time processing.",architectureImage:c,phases:[{title:"Phase 1 – Discovery & Planning (Q4 2024)",bullets:["Assessment of existing systems and processes","Definition of key performance indicators (KPIs)","Solution architecture design","Stakeholder alignment and change management planning"]},{title:"Phase 2 – Development & Testing (December 2024)",bullets:["Built AI models using Amazon SageMaker with historical fraud data","Integrated AWS services with Humano Seguros' existing IT infrastructure","Comprehensive testing with historical data including known fraud case simulation","Staff training on new systems and processes","Deployed through next-generation AI-driven DevOps pipelines, enabling automation and scalability across environments"]},{title:"Phase 3 – Pilot Launch (January 2025)",bullets:["Initial deployment processing 700 daily transactions","Real-time monitoring and AI model fine-tuning","Feedback collection from claims processing staff"]},{title:"Phase 4 – Full-Scale Deployment (February 2025)",bullets:["Scaled to handle 7,000 monthly transactions","Continuous model improvement and retraining","Integration with existing customer service systems","Enterprise-Grade AWS Security, Governance and Compliance"]}],results:{operational:["900% increase in processing capacity","50% reduction in fraud case detection time","45% of claims paid without human intervention","Processing time reduced from 9 hours to 2 minutes"],financial:["$21,000 savings in the first two months","ROI achieved within the first quarter","25% reduction in fraud investigation costs"],customerExperience:["40% fewer claims-related complaints","Faster claims resolution","More accurate fraud detection","Improved legitimate claims processing"]},authors:[{name:"Reynaldo Hidalgo",role:"Cloud Solutions Architect, AWS"},{name:"Juan Carlos Restrepo",role:"VP of Technology, Humano Seguros"},{name:"Damian Fernandez",role:"CEO, DFX5"},{name:"Gina Acosta",role:"Enterprise Sales Manager, AWS"},{name:"Barney Mancebo",role:"ICT Architecture Director, Grupo Humano"},{name:"Yaraví García de Silva",role:"Engineering Lead, DFX5"}],industry:"Health insurance",region:"Dominican Republic",techStack:[{category:"AI / ML",items:["Amazon Bedrock","SageMaker","Textract"]},{category:"Search & Data",items:["OpenSearch","S3","DynamoDB"]},{category:"Orchestration",items:["Step Functions","Lambda","EventBridge"]},{category:"Security & Compliance",items:["KMS","GuardDuty","CloudTrail"]}],pullQuote:{quote:"We went from detecting fraud in days to detecting it in seconds — without adding headcount.",author:"Juan Carlos Restrepo",role:"VP of Technology, Humano 

11Seguros"},related:["caribbean-bank-aws-modernization","rappi-ai-contact-center"]},{slug:"ansa-mcal-cloud-integration",title:"Powering Ansa McAL's Conglomerate Expansion with Cloud-Driven Integration",icon:s,accent:"from-[hsl(200,80%,55%)] to-[hsl(170,70%,50%)]",heroImage:u,outcome:"Unified systems across acquired companies on AWS, accelerating integration timelines and unlocking shared-platform synergies.",metrics:[{value:"Multi-company",label:"Unified Stack"},{value:"Faster",label:"M&A Integration"}],challenge:"Ansa McAL, a leading Caribbean conglomerate, sought a technology partner to streamline integration of acquired companies, reduce operational burden, and enable continued expansion. The goal was to unify systems and data across distributed operations, leveraging shared platforms to drive synergies.",solution:"DFX5 designed a cloud governance and integration model on AWS, consolidating shared platforms and standardizing operations across the conglomerate's distributed business units.",phases:[{title:"Phase 1 – Discovery",bullets:["Mapped systems across acquired companies","Identified shared-platform opportunities","Defined target operating model"]},{title:"Phase 2 – Foundation",bullets:["Built AWS multi-account landing zone","Standardized governance and security guardrails","Enabled shared services"]},{title:"Phase 3 – Integration",bullets:["Migrated workloads from acquisitions","Consolidated identity and access","Unified data platforms"]},{title:"Phase 4 – Optimization",bullets:["Cost optimization across accounts","Operational playbooks","Continuous improvement cadence"]}],results:{operational:["Unified governance across business units","Faster integration of new acquisitions","Standardized operations on AWS"],financial:["Reduced duplicate infrastructure costs","Shared services economies of scale"],customerExperience:["Consistent digital experience across the group"]}},{slug:"caribbean-bank-aws-modernization",title:"Transforming Banking Operations with AWS",icon:o,accent:"from-[hsl(220,80%,55%)] to-[hsl(250,70%,55%)]",heroImage:"https://images.unsplash.com/photo-1601597111158-2fceff292cdc?w=600&h=400&fit=crop",outcome:"Migrated critical banking workloads to a secure, compliant, and scalable AWS environment that meets regulatory standards.",metrics:[{value:"Secure",label:"Compliant Cloud"},{value:"Scalable",label:"Banking Platform"}],challenge:"A Caribbean financial institution needed to modernize legacy IT, improve resilience, and meet evolving regulatory requirements while controlling costs.",solution:"DFX5 led a phased migration of critical workloads to AWS — building a secure landing zone, modernizing core systems, and embedding compliance guardrails to support ongoing regulatory needs.",phases:[{title:"Phase 1 – Assessment",bullets:["Workload inventory and classification","Compliance requirements mapping","Migration wave planning"]},{title:"Phase 2 – Secure Foundation",bullets:["AWS landing zone with regulatory guardrails","Centralized logging and monitoring","Identity and key management"]},{title:"Phase 3 – Migration",bullets:["Migrated core workloads to AWS","Database modernization","Zero-downtime cutovers"]},{title:"Phase 4 – Operate",bullets:["24/7 managed services","Continuous compliance monitoring","FinOps practices"]}],results:{operational:["Secure, compliant cloud environment","Improved resilience and scalability","Modernized critical workloads"],financial:["Optimized infrastructure spend","Reduced operational overhead"],customerExperience:["Faster, more reliable banking services"]}},{slug:"bhd-digital-modernization",title:"Transforming BHD's Digital Landscape by migrating to AWS",icon:e,accent:"from-[hsl(280,70%,55%)] to-[hsl(220,80%,55%)]",heroImage:"https://images.unsplash.com/photo-1551288049-bebda4e38f71?w=600&h=400&fit=crop",outcome:"Modernized BHD's portal from legacy monolith to a scalable AWS-based architecture, enabling new digital experiences.",metrics:[{value:"Modern",label:"Architecture"},{value:"Faster",label:"Time-to-Market"}],challenge:"BHD, a leading Dominican financial services provider, faced modernizing its outdated website portal that ran on legacy tech and monolithic architecture, struggling to meet evolving customer expectations, deliver new digital experiences, and adapt to changing market demands.",solution:"DFX5 re-architected BHD's portal on AWS, decomposing the monolith into modular, scalable services and enabling a faster digital delivery cadence.",phases:[{title:"Phase 1 – Discovery",bullets:["Legacy portal assessment","Target architecture design","Migration roadmap"]},{title:"Phase 2 – Foundation",bullets:["AWS landing zone","CI/CD pipelines","Security and compliance"]},{title:"Phase 3 – Modernization",bullets:["Decomposed monolith into services","Migrated to AWS managed services","Frontend re-platform"]},{title:"Phase 4 – Launch & Optimize",bullets:["Phased rollout","Performance optimization","Ongoing modernization"]}],results:{operational:["Scalable, modern architecture on AWS","Faster release cadence","Improved reliability"],financial:["Reduced legacy maintenance costs","Optimized cloud spend"],customerExperience:["Better digital experiences for BHD customers","New capabilities delivered faster"]}},{slug:"auna-salud-genai-surgical-docs",title:"Transforming Surgical Documentation with Generative AI",icon:t,accent:"from-[hsl(160,70%,45%)] to-[hsl(190,75%,50%)]",heroImage:"https://images.unsplash.com/photo-1551601651-2a8555f1a136?w=600&h=400&fit=crop",outcome:"Deployed a GenAI voice solution on AWS that cut surgeon documentation time by 95% and boosted coding accuracy by 85%.",metrics:[{value:"95%",label:"Less Documentation Time"},{value:"85%",label:"Coding Accuracy Boost"}],challenge:"Auna Salud needed to streamline surgical documentation, which was time-consuming for surgeons and prone to coding inaccuracies that affected medical records and downstream operations.",solution:"DFX5 designed and deployed a Generative AI voice-driven documentation solution on AWS, integrating with hospital systems to capture, structure, and code surgical notes in real time.",phases:[{title:"Phase 1 – Discovery",bullets:["Surgeon workflow analysis","Use case definition","Architecture design"]},{title:"Phase 2 – Build",bullets:["GenAI voice pipeline on AWS","Integration with EMR systems","Coding model fine-tuning"]},{title:"Phase 3 – Pilot",bullets:["Pilot with surgical teams","Accuracy validation","Workflow refinement"]},{title:"Phase 4 – Scale",bullets:["Hospital-wide rollout","Continuous model improvement","Ongoing managed services"]}],results:{operational:["95% reduction in surgeon documentation time","85% improvement in coding accuracy","Improved medical records quality"],financial:["Operational efficiency gains","Reduced administrative overhead"],customerExperience:["More time for patient care","Higher-quality clinical documentation"]}},{slug:"sdlc-agentic-ai-payments",title:"Accelerating the Enterprise SDLC with Multi-Agent Agentic AI on AWS",icon:t,accent:"from-[hsl(250,70%,55%)] to-[hsl(200,80%,55%)]",heroImage:m,outcome:"Delivered a multi-agent Agentic AI platform on Amazon Bedrock and AgentCore Runtime — four production-grade agents shipped on plan, targeting a 70% reduction in SDLC planning time and 40% development-velocity uplift.",metrics:[{value:"70%",label:"Planning Time Reduction"},{value:"40%",label:"Dev Velocity Uplift"}],challenge:"A leading Latin American payments-technology company was constrained by an SDLC that had become a structural drag on time-to-market: two-month planning cycles for every effort regardless of size, requirements analysis lacking technical depth, fragmented documentation across four disconnected systems, insufficient code-review time, no comprehensive impact analysis on the regulated payments platform, and shared development environments that throttled parallel work.",solution:"DFX5 designed and deployed a multi-agent Agentic AI platform — an orchestrator hub plus three specialized agents with a documented 12-agent roadmap — built on Amazon Bedrock (Claude Sonnet and Haiku, Titan Embeddings v2), Amazon Bedrock AgentC

11ore Runtime, AWS Lambda, Bedrock Knowledge Base with S3 Vectors, API Gateway, EventBridge, DynamoDB, and the Strands Agents framework. Agentic workloads run in an AWS account strictly segregated from PCI-DSS-bound payment-processing accounts, with PrivateLink keeping model-invocation traffic off the public internet and human-in-the-loop approval gating every agent output.",phases:[{title:"Sprint 1 – Orchestrator Agent",bullets:["Stateful hub on AgentCore Runtime that classifies intent and routes work","Multi-turn refinement powered by AgentCore Memory","Observability and prompt-as-code patterns from Sprint 0"]},{title:"Sprint 2 – Impact-Analysis Agent",bullets:["Event-driven analysis of proposed changes against the payments platform","Surfaces downstream effects before code is written","Serverless Lambda execution with right-sized memory from CloudWatch metrics"]},{title:"Sprint 3 – Document-Analysis Agent",bullets:["High-throughput RAG over previously scattered documentation corpus","Bedrock Knowledge Base with Amazon S3 Vectors for managed retrieval","Unified access to requirements and technical documentation"]},{title:"Sprint 4 – Epics-and-Stories Agent",bullets:["Generation of structured epics and user stories on AgentCore Runtime","Multi-turn human refinement before commit to backlog","Automated CI/CD promotion via CloudFormation, CodePipeline, CodeBuild, CodeDeploy"]}],results:{operational:["Four production-grade agents shipped on plan — one per sprint","100% end-to-end test pass rate validated via automated CI/CD gates","Agents promoted to the customer's quality (UAT) environment"],financial:["Per-agent foundation-model selection optimizes cost-to-quality","Right-sized Lambda memory and on-demand DynamoDB billing","Amazon S3 Vectors avoids self-managed vector cluster overhead"],customerExperience:["Targeted 70% reduction in SDLC planning time","Targeted 40% uplift in development velocity","Targeted 95% accuracy on impact and risk analysis"]},industry:"Financial Services — Payments Technology",region:"Latin America",techStack:[{category:"AI / ML",items:["Amazon Bedrock","AgentCore Runtime","Claude Sonnet & Haiku","Titan Embeddings v2"]},{category:"Compute & Orchestration",items:["AWS Lambda","API Gateway","EventBridge","DynamoDB"]},{category:"Data & RAG",items:["Bedrock Knowledge Base","Amazon S3 Vectors","Strands Agents"]},{category:"Security & Governance",items:["Cognito","WAF","KMS","Secrets Manager","PrivateLink","Control Tower"]},{category:"DevOps",items:["CloudFormation","CodePipeline","CodeBuild","CodeDeploy"]}],pullQuote:{quote:"Every agent output is a draft for explicit human approval — no agent action is applied autonomously to our backlog, source control, or documentation.",author:"Engineering Leadership",role:"Payments Technology Customer"},related:["rappi-ai-contact-center","humano-seguros-fraud-detection"]},{slug:"mining-maintenance-agentic-ai",title:"Transforming Mining Maintenance Operations with Agentic AI on AWS",icon:r,accent:"from-[hsl(35,85%,50%)] to-[hsl(15,80%,45%)]",heroImage:g,outcome:"Replaced a manual radio-based information-retrieval process with an Agentic AI assistant on AWS — cutting equipment information access time from 15 minutes to under 1 minute for ~400 daily field users at ~89% lower operational cost.",metrics:[{value:"93%",label:"Faster Info Access"},{value:"89%",label:"Lower Op Cost"}],challenge:"At a remote high-altitude copper-and-zinc complex processing over 100,000 tonnes of ore per day, maintenance technicians needed equipment information siloed across multiple SAP-based systems accessible only to specialized office staff. Field crews faced an average 15-minute wait per request via radio call, generating 2,000–4,000 daily requests, lost productivity, and elevated risk of unplanned downtime — which can cost $50,000–$100,000 per hour on critical mining assets.",solution:"DFX5 designed and deployed an Agentic AI chatbot platform on AWS — Amazon Bedrock with Claude models for Spanish natural-language understanding and agentic reasoning, AWS Lambda for autonomous tool routing across 8 maintenance APIs, AWS Step Functions for workflow orchestration, Amazon API Gateway with JWT validation and WAF protection, and Amazon S3 as the data lake. The architecture integrates with 8 existing SAP and proprietary maintenance APIs under strict read-only access — the assistant can never autonomously modify production data — and runs serverless with zero cost at rest during off-hours.",phases:[{title:"Phase 1 – Discovery",bullets:["Mapped the field-maintenance information journey end to end","Identified 8 back-end query categories across SAP systems of record","Defined read-only integration as a hard safety boundary"]},{title:"Phase 2 – Agentic Orchestration Build",bullets:["Intent classification and confidence thresholding (≥ 0.7) in Lambda","Autonomous tool routing across 8 maintenance APIs","Circuit-breaker retry logic and serverless zero-cost-at-rest design"]},{title:"Phase 3 – Secure Integration",bullets:["Connected 8 maintenance APIs via proxy authentication, read-only","API Gateway with JWT validation and 100 req/min rate limiting","AWS WAF for OWASP Top 10 protection on the field-accessible endpoint"]},{title:"Phase 4 – Rollout & Adoption",bullets:["Spanish-language assistant deployed to ~400 maintenance users","Real-time monitoring of usage and intent-classification accuracy","Clarification prompts automatically triggered for low-confidence queries"]}],results:{operational:["~467 technician-hours per day freed for higher-value maintenance work","92% intent-classification accuracy across 8 query categories","99.7% availability with sub-second P95 latency"],financial:["~$280/month total operational cost on serverless AWS","~89% savings versus traditional infrastru

11cture","~$250,000/month in technician capacity redirected to higher-value work"],customerExperience:["61% user adoption in the first month (245 of 400 target users)","Sustained engagement of 5–10 queries per user per day","Spanish self-service from field locations with no specialized IT assistance"]},industry:"Mining & Natural Resources",region:"South America",techStack:[{category:"AI / ML",items:["Amazon Bedrock","Anthropic Claude","Spanish NLU"]},{category:"Compute & Orchestration",items:["AWS Lambda (Node.js)","Step Functions","API Gateway"]},{category:"Data & Storage",items:["Amazon S3","Data Lake"]},{category:"Security",items:["AWS WAF","JWT Auth","IAM","Rate Limiting"]},{category:"Integration",items:["8 SAP / Maintenance APIs","Proxy Auth","Read-Only Access"]}],pullQuote:{quote:"The assistant has read-only access to every system it queries — agentic autonomy with a hard safety boundary, and zero cost at rest when the mine sleeps.",author:"Maintenance Operations Leadership",role:"Mining & Natural Resources Customer"},related:["sdlc-agentic-ai-payments","humano-seguros-fraud-detection"]}];export{h as c};