[{"data":1,"prerenderedAt":1665},["ShallowReactive",2],{"uc-reg-iso-42001":3},{"regulation":4,"includeUnpublished":11,"indexable":12,"useCases":13},{"id":5,"label":6,"issuer":7,"region":8,"url":9,"description":10},"iso-42001","ISO/IEC 42001","ISO and IEC","global","https://www.iso.org/standard/81230.html","The international management system standard for AI.",false,true,[14,55,79,101,119,148,167,191,207,226,245,268,291,312,328,342,355,368,385,398,416,427,441,460,475,487,504,521,534,548,568,582,601,617,631,646,658,671,686,704,722,736,752,768,780,793,806,821,837,851,863,876,888,900,915,929,943,959,972,989,1003,1016,1029,1039,1051,1066,1085,1100,1114,1130,1149,1161,1173,1185,1196,1210,1222,1233,1244,1260,1271,1284,1296,1310,1324,1337,1351,1362,1373,1387,1400,1415,1434,1448,1460,1476,1490,1504,1516,1528,1538,1550,1561,1577,1589,1602,1615,1626,1640,1653],{"slug":15,"title":16,"shortTitle":17,"definition":18,"status":19,"industries":20,"functions":22,"patterns":25,"audience":30,"autonomy":31,"adoptionStage":32,"segment":33,"evidenceCount":34,"publicEvidenceCount":34,"organizations":35,"bestGrade":41,"headline":42,"lastVerified":52,"indexable":12,"euAiActTier":53,"euAiActBasis":54},"autonomous-network-operations","Agentic AI for autonomous, intent based network operations","Autonomous network operations","AI agents that run closed loops over a telecom network: they take an intent from the operator (for example a latency or availability target for a service), observe the network, diagnose deviations and execute corrective actions across radio, transport and core, within guardrails set by engineers and with human approval for major changes.","published",[21],"telecommunications",[23,24],"network-operations","it-and-engineering",[26,27,28,29],"agentic-workflow","anomaly-detection","prediction-and-scoring","classification-and-routing","back-office","supervised-agent","emerging","network",6,[36,37,38,39,40],"Deutsche Telekom","du","KDDI","stc Group","Telstra","B",{"kpi":43,"label":44,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":49,"qualifier":50,"claimant":51,"organization":36,"vendorReported":11},"processing-time-reduction","Cycle time reduction","percent",1,0,"reported",95,"at-least","organization","2026-09-26","context-dependent","Annex III point 2 lists AI systems intended as safety components in the management and operation of critical digital infrastructure as high risk; Recital 55 ties this to the digital infrastructure in the Annex to Directive (EU) 2022/2557, which includes providers of public electronic communications networks. Recital 55 defines such safety components as systems that directly protect the physical integrity of the infrastructure or the health and safety of persons and property, and excludes components used solely for cybersecurity. Loops that only optimise performance or capacity are usually not safety components, but a loop that protects physical integrity or life safety services can be, so operators should assess each closed loop and document the outcome.",{"slug":56,"title":57,"shortTitle":58,"definition":59,"status":19,"industries":60,"functions":63,"patterns":64,"audience":65,"autonomy":31,"adoptionStage":66,"evidenceCount":67,"publicEvidenceCount":67,"organizations":68,"bestGrade":71,"headline":72,"lastVerified":76,"indexable":12,"euAiActTier":77,"euAiActBasis":78},"cloud-cost-optimization-agent","AI agent for cloud cost optimization and FinOps","Cloud cost optimization agent","An AI agent that continuously reads an organization's cloud usage and billing data, uses machine learning to separate normal spend from waste, and either rightsizes resources and buys a mix of committed capacity matched to forecast usage on its own within set limits, or proposes higher risk changes for an engineer to approve.",[61,62],"cross-industry","technology",[24],[27,28,26],"employee-facing","early-adopters",2,[69,70],"Akamai Technologies","VERMEG","C",{"kpi":73,"label":74,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":75,"qualifier":50,"claimant":51,"organization":69,"vendorReported":11},"cost-reduction","Cost reduction",40,"2026-09-28","minimal","An internal tool that optimizes infrastructure spend and makes no decision about a natural person, so the default case falls outside Annex III. The relevant Annex III entry to check against is point 2, AI safety components in the management and operation of critical digital infrastructure: a cost agent stays outside it as long as its policy keeps it to cost actions (rightsizing, commitment purchases, idle cleanup) rather than acting as a safety component of the infrastructure itself. The Akamai deployment on this page shows the scope can extend to core production infrastructure, so an operator should confirm this against its own policy rather than assume it.",{"slug":80,"title":81,"shortTitle":82,"definition":83,"status":19,"industries":84,"functions":86,"patterns":88,"audience":65,"autonomy":90,"adoptionStage":66,"evidenceCount":67,"publicEvidenceCount":67,"organizations":91,"bestGrade":71,"headline":94,"lastVerified":76,"indexable":12,"euAiActTier":77,"euAiActBasis":100},"data-quality-monitoring-agent","AI agent for data quality monitoring and observability","Data quality monitoring agent","An AI agent that watches data pipelines and tables continuously, uses machine learning to learn the normal pattern of freshness, volume, schema and distribution for each one, flags anomalies before they reach a dashboard or a downstream model, and traces the lineage back to the change that caused them so an engineer can fix the source, not just the symptom.",[61,62,85],"retail-and-ecommerce",[24,87],"analytics-and-reporting",[27,29,89],"summarization","assist",[92,93],"Contentsquare","SeatGeek",{"kpi":95,"label":96,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":97,"qualifier":98,"claimant":99,"organization":93,"vendorReported":12},"productivity-gain","Productivity gain",50,"exact","vendor","An internal data engineering tool that flags anomalies in pipelines and tables; it is not a use listed in Annex III and makes no decision about a natural person. If the monitored data feeds a high risk system, such as a credit or employment decision, the AI Act obligations attach to that downstream system, not to this monitoring layer.",{"slug":102,"title":103,"shortTitle":104,"definition":105,"status":19,"industries":106,"functions":107,"patterns":109,"audience":65,"autonomy":111,"adoptionStage":66,"evidenceCount":67,"publicEvidenceCount":67,"organizations":112,"bestGrade":71,"headline":115,"lastVerified":76,"indexable":12,"euAiActTier":117,"euAiActBasis":118},"performance-review-drafting-agent","AI agent for drafting employee performance reviews","Performance review drafting","An assistant that gathers an employee's work history, goals and peer feedback from the systems a manager already uses, and drafts a first version of the performance review for the manager to edit, rewrite or reject, so the manager starts from a grounded summary instead of a blank form and a stack of six months of context to recall from memory.",[61,62],[108],"human-resources",[110,89,26],"content-generation","copilot",[113,114],"Case Status","Rho",{"kpi":43,"label":44,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":116,"qualifier":98,"claimant":99,"organization":113,"vendorReported":12},84,"high","Annex III point 4(b) lists AI systems intended to monitor and evaluate the performance and behaviour of workers as high risk. Synthesising an employee's work history and feedback into a performance evaluation is very plausibly profiling of a natural person under GDPR Article 4(4), which expressly covers analysing or predicting a person's \"performance at work\". Article 6(3)'s last subparagraph makes an Annex III system high risk regardless of the derogations whenever it performs such profiling, so a tool built this way is high risk by default however much the manager edits the output. The derogations in Article 6(3), including a narrow procedural task or improving the result of a previously completed human activity, do not fit drafting an evaluation from scratch; the closest is point (d), a preparatory task ahead of a human assessment, which only has a chance of applying to a design that avoids profiling altogether, for example one that only surfaces raw facts without synthesising a judgement. Where that derogation is argued, the documentation duty under Article 6(4) falls on the provider of the system, and only on the deploying organization when it builds the tool itself. Because the tool is high risk by default, Article 26(7) requires informing affected workers and their representatives before it is put into use in the workplace, whatever the tool's output is used for; using the same system's output directly in pay, promotion or termination decisions removes any doubt and triggers the full high risk regime. Annex III's high risk obligations apply from 2 December 2027.",{"slug":120,"title":121,"shortTitle":122,"definition":123,"status":19,"industries":124,"functions":126,"patterns":129,"audience":133,"autonomy":31,"adoptionStage":66,"segment":127,"evidenceCount":134,"publicEvidenceCount":134,"organizations":135,"bestGrade":41,"headline":141,"lastVerified":145,"indexable":12,"euAiActTier":146,"euAiActBasis":147},"claims-first-notice-of-loss-agent","AI agent for first notice of loss claims intake","First notice of loss agent","An AI agent that takes the first notice of loss from a policyholder by phone, chat or app, identifies the policy, collects the facts of the incident and the evidence the claim type needs, opens the claim in the claims system and tells the customer what happens next, handing complex, injured or vulnerable claimants to a human handler.",[125],"insurance",[127,128],"claims","customer-service",[130,131,26,132],"conversational-agent","voice-agent","document-processing","customer-facing",5,[136,137,138,139,140],"DOMCURA","Hippo","Lemonade","Progressive","Travelers",{"kpi":142,"label":143,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":144,"qualifier":98,"claimant":99,"organization":136,"vendorReported":12},"accuracy","Accuracy",90,"2026-09-27","limited","A customer facing intake agent must be designed so that people know they are interacting with AI (Article 50(1)). Claims intake and claims handling are not listed in Annex III: point 5(c) covers risk assessment and pricing in life and health insurance, not claims. One design choice changes this: detecting distress by inferring emotions from the caller's voice is emotion recognition based on biometric data, which is high risk under Annex III point 1(c) and needs disclosure under Article 50(3). Detecting vulnerability from what the caller says does not. The limited tier assumes that design: every handover signal on this page (injury, distress, anger, vulnerability) is detected from the words of the conversation, and inferring emotions from the voice itself is out of scope.",{"slug":149,"title":150,"shortTitle":151,"definition":152,"status":19,"industries":153,"functions":156,"patterns":159,"audience":65,"autonomy":31,"adoptionStage":66,"segment":160,"evidenceCount":161,"publicEvidenceCount":67,"organizations":162,"bestGrade":71,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":166},"fraud-alert-triage","AI agent for fraud alert triage","Fraud alert triage","An AI agent that works the fraud alert queue behind the scenes as the analyst's first pass, without contacting the customer: it enriches each alert with customer, device and payment context, closes clear false positives under documented rules, merges duplicates, and routes genuine risk to an analyst with a drafted rationale.",[154,155],"banking","payments",[157,158],"fraud-prevention","operations",[26,29,89,28],"middle-office",3,[163,164],"Coast","SEB",null,"Internal triage of fraud alerts is not listed in Annex III, and point 5(b) explicitly excludes fraud detection from the high risk creditworthiness category. Article 50(1) covers any system that interacts directly with people, analysts included, but it does not apply where the use of AI is obvious to a reasonably well informed user, as it is in an internal analyst tool; the marking duties for generated content in Article 50(2) sit with the provider. Reassess if its output feeds credit decisions. Decisions that affect customers remain subject to GDPR and consumer protection rules.",{"slug":168,"title":169,"shortTitle":170,"definition":171,"status":19,"industries":172,"functions":174,"patterns":175,"audience":65,"autonomy":31,"adoptionStage":177,"evidenceCount":178,"publicEvidenceCount":34,"organizations":179,"bestGrade":41,"headline":186,"lastVerified":145,"indexable":12,"euAiActTier":146,"euAiActBasis":190},"it-service-desk-resolution-agent","AI agent for IT service desk resolution","IT service desk resolution","An AI agent in Microsoft Teams, Slack or the intranet that takes the high volume IT support queue, such as password and MFA resets, account unlocks, VPN, device and software requests, and resolves common requests by acting in the identity and IT service management systems, handing the rest to the right resolver group with the context attached.",[61,154,62,85,173],"healthcare",[24,158],[130,26,176,29],"rag-knowledge-assistant","mainstream",8,[180,181,182,183,184,185],"7-Eleven Vietnam","Bank of America","Equinix","IBM","Mercari US","Vituity",{"kpi":187,"label":188,"unit":45,"n":67,"nUpTo":47,"kind":48,"value":189,"qualifier":98,"claimant":99,"organization":184,"vendorReported":12},"employee-adoption","Employee adoption",94,"Article 50(1) requires an assistant that talks with people to make clear they are interacting with AI, unless that is obvious from the context. It is not listed in Annex III. The agent does allocate work, but it routes tickets to resolver and assignment groups based on the content of the request, not to individual workers based on their behaviour or personal traits or characteristics, so Annex III point 4(b) does not apply. It also does not decide on recruitment, promotion, credit or access to essential services. Any use that assigns work to individual analysts, or monitors and evaluates them from their behaviour or performance (including through the agent's logs), would need its own assessment.",{"slug":192,"title":193,"shortTitle":194,"definition":195,"status":19,"industries":196,"functions":198,"patterns":201,"audience":65,"autonomy":111,"adoptionStage":66,"segment":202,"evidenceCount":161,"publicEvidenceCount":161,"organizations":203,"bestGrade":41,"headline":165,"lastVerified":52,"indexable":12,"euAiActTier":53,"euAiActBasis":206},"source-of-wealth-diligence","AI agent for source of wealth due diligence in private banking","Source of wealth diligence","An AI agent that reads a prospective private client's documents, extracts and corroborates how their wealth was built, checks plausibility against benchmarks and external sources, and drafts the source of wealth and enhanced due diligence narrative for the relationship manager and compliance analyst, who decide on the risk rating and the relationship.",[197,154],"wealth-and-asset-management",[199,200],"onboarding-and-kyc","financial-crime-compliance",[132,26,110,89],"front-office",[204,205],"Bank of Singapore","Deutsche Bank","Anti money laundering due diligence is not listed in Annex III, so an assistant that drafts source of wealth reports for a human decision is not high risk by default. It becomes high risk if it adds remote biometric identification of the client (Annex III point 1(a); verification that only confirms a claimed identity is excluded) or feeds an assessment of a natural person's creditworthiness, for example for lending to the client (Annex III point 5(b)). GDPR Article 22 on solely automated decisions applies if it ever refused a client on its own.",{"slug":208,"title":209,"shortTitle":210,"definition":211,"status":19,"industries":212,"functions":214,"patterns":215,"audience":133,"autonomy":31,"adoptionStage":32,"segment":127,"evidenceCount":161,"publicEvidenceCount":67,"organizations":217,"bestGrade":41,"headline":220,"lastVerified":145,"indexable":12,"euAiActTier":146,"euAiActBasis":225},"travel-insurance-claims-and-assistance-agent","AI agent for travel insurance claims and assistance","Travel insurance claims and assistance","An AI agent that helps insured travellers around the clock and in their own language: it answers cover questions, takes claims for delays, cancellations, lost baggage and medical costs, reads the receipts and certificates they upload, settles simple claims within set limits, and connects medical emergencies and complex cases to the assistance team at once.",[125,213],"travel-and-hospitality",[127,128],[130,131,132,26,216],"translation",[218,219],"Allianz Partners","General Insurance Association of Singapore",{"kpi":221,"label":222,"unit":45,"n":47,"nUpTo":46,"kind":48,"value":223,"qualifier":224,"claimant":51,"organization":218,"vendorReported":11},"automation-rate","Automation rate",70,"up-to","A customer facing agent must disclose that it is AI (Article 50), unless this is obvious from the context. Travel insurance claims handling is not listed in Annex III; point 5(c) covers risk assessment and pricing in life and health insurance, not the handling of claims. Handing a traveller who reports a medical emergency to the assistance team is not the classification of emergency calls or the patient triage in point 5(d), as long as the agent only hands over and does not set medical priorities. Claim decisions based solely on automated processing are subject to GDPR Article 22 (and its UK equivalent), and medical data is special category data under Article 9.",{"slug":227,"title":228,"shortTitle":229,"definition":230,"status":19,"industries":231,"functions":232,"patterns":234,"audience":65,"autonomy":111,"adoptionStage":177,"evidenceCount":161,"publicEvidenceCount":161,"organizations":236,"bestGrade":41,"headline":240,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":244},"ambient-clinical-documentation","AI ambient scribe for clinical documentation","Ambient clinical documentation","An AI scribe that listens, with the patient's consent, to the conversation between a clinician and a patient and drafts the clinical note, and often the letter or after visit summary, for the clinician to review, edit and sign in the health record. It documents; it does not diagnose or decide on treatment.",[173],[158,233],"knowledge-management",[235,89,110],"speech-analytics",[237,238,239],"Great Ormond Street Hospital for Children NHS Foundation Trust","Kaiser Permanente","US Department of Veterans Affairs, Veterans Health Administration",{"kpi":241,"label":242,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":243,"qualifier":98,"claimant":51,"organization":237,"vendorReported":11},"handling-time-reduction","Handling time reduction",8.2,"A scribe that only transcribes and summarises for a clinician to review is not listed in Annex III and is usually minimal risk, although the provider of a system that generates text can still owe the Article 50(2) duty to mark output as AI generated, unless an exception such as an assistive function for standard editing applies. If the product qualifies as medical device software under the EU Medical Device Regulation and needs a notified body assessment, for example because it suggests diagnoses or treatment, it becomes high risk under Article 6(1) and Annex I. Health data in audio and notes falls under GDPR Article 9 in every case.",{"slug":246,"title":247,"shortTitle":248,"definition":249,"status":19,"industries":250,"functions":252,"patterns":255,"audience":133,"autonomy":111,"adoptionStage":66,"evidenceCount":256,"publicEvidenceCount":256,"organizations":257,"bestGrade":41,"headline":265,"lastVerified":52,"indexable":12,"euAiActTier":53,"euAiActBasis":267},"benefits-eligibility-and-application-assistant","AI assistant for benefits eligibility questions and applications","Benefits eligibility and application assistant","An AI assistant that helps people understand which public benefits and grants may apply to them, explains the rules and documents in plain language, guides them through the application and checks it for completeness, while the eligibility decision stays with the agency's rules and caseworkers.",[251],"government",[253,254,128],"citizen-services","case-management",[130,176,131,132],7,[258,259,260,261,262,263,264],"Department for Work and Pensions","Federal Student Aid (U.S. Department of Education)","Federal Emergency Management Agency","Gemeente Nissewaard","Leeds City Council","Région Provence-Alpes-Côte d'Azur (Région Sud)","YoungWilliams",{"kpi":142,"label":143,"unit":45,"n":67,"nUpTo":47,"kind":48,"value":266,"qualifier":98,"claimant":51,"organization":258,"vendorReported":11},97,"Annex III point 5(a) makes AI high risk when it is used by or on behalf of public authorities to evaluate the eligibility of natural persons for essential public assistance benefits and services, or to grant, reduce, revoke or reclaim them. An assistant that only explains rules and guides applications carries the Article 50 transparency duties (limited risk); one that screens or scores eligibility falls under point 5(a), and a public body deploying it must carry out a fundamental rights impact assessment first (Article 27).",{"slug":269,"title":270,"shortTitle":271,"definition":272,"status":19,"industries":273,"functions":274,"patterns":275,"audience":133,"autonomy":31,"adoptionStage":177,"evidenceCount":276,"publicEvidenceCount":277,"organizations":278,"bestGrade":41,"headline":288,"lastVerified":145,"indexable":12,"euAiActTier":146,"euAiActBasis":290},"citizen-information-assistant","AI assistant for citizen information and government services","Citizen information assistant","An AI assistant that answers residents' and businesses' questions about government services in plain language, grounded only in official guidance with links to the source, points them to the right online service or office, and hands anything personal, urgent or outside its content to a human with the context attached.",[251],[253,128,233],[176,130,131,29],11,9,[279,280,281,282,283,284,285,286,287],"Abu Dhabi Government (TAMM)","Driver and Vehicle Licensing Agency","Estonian Information System Authority (RIA)","Foreign, Commonwealth and Development Office","Gemeente Tilburg","Government Digital Service","Government of the City of Buenos Aires","Madrid Destino","Montgomery County Government",{"kpi":142,"label":143,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":289,"qualifier":50,"claimant":51,"organization":282,"vendorReported":11},76,"An information assistant must tell people they are interacting with AI (Article 50). It is not high risk as long as it does not evaluate eligibility for public assistance benefits or services (Annex III point 5(a)); an assistant that starts to pre assess eligibility should be reassessed.",{"slug":292,"title":293,"shortTitle":294,"definition":295,"status":19,"industries":296,"functions":297,"patterns":298,"audience":133,"autonomy":90,"adoptionStage":66,"segment":300,"evidenceCount":301,"publicEvidenceCount":301,"organizations":302,"bestGrade":41,"headline":307,"lastVerified":145,"indexable":12,"euAiActTier":146,"euAiActBasis":311},"developer-api-integration-assistant","AI assistant for developers integrating a company's APIs","API integration assistant","An AI assistant on a developer portal and in its documentation that answers integration questions, recommends the right endpoints, helps debug connections and generates sample calls, grounded in the API catalogue, reference docs and test material, so clients and partners integrate faster with fewer support tickets.",[61,154,155,62],[24,128,199],[176,130,299],"code-generation","specialized-businesses",4,[303,304,305,306],"CircleCI","Mapbox","monday.com","U.S. Bank",{"kpi":308,"label":309,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":310,"qualifier":98,"claimant":51,"organization":304,"vendorReported":11},"contact-deflection","Contact deflection",30,"A chatbot that interacts with developers must disclose that it is AI (Article 50). Code generation for integration is not listed in Annex III.",{"slug":313,"title":314,"shortTitle":315,"definition":316,"status":19,"industries":317,"functions":318,"patterns":320,"audience":133,"autonomy":31,"adoptionStage":66,"segment":202,"evidenceCount":34,"publicEvidenceCount":161,"organizations":322,"bestGrade":71,"headline":325,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":327},"digital-onboarding-assistant","AI assistant for digital account onboarding and KYC","Digital onboarding","A customer facing AI assistant that guides a new applicant, a person or a small merchant, through a digital account, card or relationship application: it collects and checks identity and supporting documents, orchestrates the know your customer and anti money laundering checks, prefills what it can and sends only the unclear cases to a human reviewer with a summary. The ownership research for complex corporate clients is a separate back office job.",[154,155,197],[199,319,128],"sales",[130,132,321,26],"computer-vision",[323,205,324],"Albo","M-DAQ Global",{"kpi":95,"label":96,"unit":326,"n":46,"nUpTo":47,"kind":48,"value":310,"qualifier":98,"claimant":99,"organization":324,"vendorReported":12},"multiplier","The conversational assistant falls under the Article 50 transparency duty. Biometric verification whose sole purpose is to confirm that a person is who they claim to be is excluded from the Annex III biometric category. The system becomes high risk when the same journey assesses creditworthiness or a credit score of a natural person, for example for a credit card or overdraft (Annex III point 5(b)).",{"slug":329,"title":330,"shortTitle":331,"definition":332,"status":19,"industries":333,"functions":335,"patterns":336,"audience":65,"autonomy":31,"adoptionStage":66,"evidenceCount":301,"publicEvidenceCount":161,"organizations":337,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":341},"employee-onboarding-assistant","AI assistant for employee onboarding","Employee onboarding assistant","An assistant that guides each new employee from signed contract through the first months: it answers first week questions in plain language, tracks the personal onboarding checklist, triggers the paperwork, equipment, access and training steps in the systems that own them, and keeps the manager and HR informed of what is still open.",[61,251,334,173],"professional-services",[108,233],[130,176,26],[338,339,340],"American Addiction Centers","KPMG","U.S. Department of Agriculture","Answering onboarding questions and orchestrating provisioning is limited risk: under Article 50(1) the assistant must be designed so that employees are told they are interacting with AI, unless that is obvious. It becomes high risk under Annex III point 4(b) if it is used to make decisions on the terms or termination of the work relationship, to allocate tasks based on individual behaviour or personal traits, or to monitor and evaluate new hires' performance or behaviour, for example to judge probation.",{"slug":343,"title":344,"shortTitle":345,"definition":346,"status":19,"industries":347,"functions":348,"patterns":350,"audience":65,"autonomy":111,"adoptionStage":32,"segment":202,"evidenceCount":67,"publicEvidenceCount":67,"organizations":351,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":354},"goal-based-financial-planning-assistant","AI assistant for goal based financial planning","Goal based planning","An AI assistant that turns a client's goals into projections and what if scenarios using a rules based planning engine, explains the trade offs in plain language and prepares the plan for an advisor to validate, with every assumption disclosed and reproducible.",[197,154],[319,128,349],"product-and-pricing",[130,110,26],[352,353],"CIMB Niaga","Vanguard","Planning support for advisors is not listed in Annex III. A client facing version must disclose that the client is talking to AI (Article 50). It becomes high risk if it is used to assess the creditworthiness of individuals (Annex III point 5(b)) or for risk assessment and pricing of life or health insurance for individuals (Annex III point 5(c)).",{"slug":356,"title":357,"shortTitle":358,"definition":359,"status":19,"industries":360,"functions":361,"patterns":362,"audience":65,"autonomy":31,"adoptionStage":66,"evidenceCount":134,"publicEvidenceCount":301,"organizations":363,"bestGrade":41,"headline":365,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":367},"hr-and-policy-assistant","AI assistant for HR and policy questions","HR and policy assistant","An employee self service assistant that answers questions on leave, pay and tax forms, benefits, expenses, travel and conduct policies from the organization's own HR documents, personalized to the employee's country and role, and starts simple HR transactions such as leave requests or employment letters in the HR system.",[61,154,62,173],[108,233],[176,130,26],[181,183,364,185],"Turing",{"kpi":187,"label":188,"unit":45,"n":67,"nUpTo":47,"kind":48,"value":366,"qualifier":98,"claimant":51,"organization":183,"vendorReported":11},99,"Answering policy questions and starting routine requests is limited risk, with the Article 50 duty to disclose AI. It becomes high risk under Annex III point 4 if it is used to make or support decisions on recruitment, promotion, termination, allocating tasks based on individual behaviour or personal traits, or the monitoring and evaluation of workers; an employer deploying it then must also inform workers' representatives and the affected workers before use (Article 26(7)). Sensitive topic detection should work on what the employee writes: inferring emotions of people in the workplace from biometric data such as voice or facial expressions is prohibited under Article 5(1)(f), except for medical or safety reasons.",{"slug":369,"title":370,"shortTitle":371,"definition":372,"status":19,"industries":373,"functions":374,"patterns":375,"audience":65,"autonomy":90,"adoptionStage":66,"segment":377,"evidenceCount":134,"publicEvidenceCount":134,"organizations":378,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":384},"insurance-broker-and-agent-assistant","AI assistant for insurance brokers and agents","Broker and agent assistant","An AI assistant for tied agents, independent brokers, advisors and the insurer's own distribution staff that answers product, underwriting and process questions from approved sources, prepares personalized customer engagement and follow ups, validates and prioritizes leads, and drafts meeting notes and emails, so producers spend more time with customers.",[125],[319,233],[176,376,110,89],"recommendation-and-personalization","distribution",[379,380,381,382,383],"Manulife","Prudential plc","Sun Life","Waterdrop","Zurich Insurance Group","An employee facing assistant for knowledge answers and drafting is not listed in Annex III and is minimal risk. A lead qualification agent that talks to customers must tell them they are dealing with AI (Article 50). Using performance insights to monitor and evaluate individual agents, or to allocate leads based on their behaviour or traits, is high risk under Annex III point 4(b), and any component that does risk assessment or pricing of life or health insurance for individuals is high risk under Annex III point 5(c).",{"slug":386,"title":387,"shortTitle":388,"definition":389,"status":19,"industries":390,"functions":391,"patterns":394,"audience":65,"autonomy":111,"adoptionStage":32,"segment":202,"evidenceCount":67,"publicEvidenceCount":67,"organizations":395,"bestGrade":41,"headline":165,"lastVerified":52,"indexable":12,"euAiActTier":53,"euAiActBasis":397},"suitability-assessment-assistant","AI assistant for investment suitability assessment and reports","Suitability assessment","An AI assistant that checks whether a proposed product or portfolio fits a client's risk tolerance, objectives, knowledge, experience and financial situation against the firm's rules, flags mismatches, and drafts the suitability rationale and report for the advisor to confirm, while hard rule failures are decided by deterministic checks, not by the model.",[197,154],[392,319,393],"regulatory-compliance","risk-management",[26,110,29],[396,353],"Morgan Stanley","Investment suitability assessment is not listed in Annex III, so the tier depends on design. It becomes high risk where the same system assesses creditworthiness, for example for lending against a portfolio (Annex III point 5(b)). MiFID II suitability duties apply regardless of the AI Act tier.",{"slug":399,"title":400,"shortTitle":401,"definition":402,"status":19,"industries":403,"functions":405,"patterns":409,"audience":65,"autonomy":111,"adoptionStage":66,"evidenceCount":34,"publicEvidenceCount":134,"organizations":410,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":415},"procurement-contract-review","AI assistant for procurement and supplier contract review","Procurement and contract review","An assistant for procurement and vendor management that reads supplier contracts and proposals, extracts the key terms, flags deviations from the organization's standard positions, drafts requests for proposal and evaluation matrices, and prepares negotiation positions, with a procurement or legal owner approving every conclusion.",[61,154,251,85,404],"manufacturing",[406,407,408],"procurement","legal","finance-and-accounting",[132,176,110,26],[411,412,413,414],"General Services Administration","Administration for Children and Families","Internal Revenue Service","Walmart","Contract review and sourcing are not among the Annex III high risk uses, so an internal assistant that makes no decisions about natural persons is minimal risk (with the Article 4 AI literacy duty). If a negotiation bot chats directly with supplier staff, Article 50(1) applies and it must tell them they are dealing with an AI system, unless that is obvious from the context. Public authorities using AI in procurement should still check national public procurement rules on transparency and equal treatment of bidders.",{"slug":417,"title":418,"shortTitle":419,"definition":420,"status":19,"industries":421,"functions":422,"patterns":423,"audience":65,"autonomy":111,"adoptionStage":32,"segment":300,"evidenceCount":46,"publicEvidenceCount":46,"organizations":424,"bestGrade":41,"headline":165,"lastVerified":52,"indexable":12,"euAiActTier":77,"euAiActBasis":426},"shariah-compliance-screening","AI assistant for Shariah compliance screening and review","Shariah compliance screening","An AI assistant that screens Islamic financing contracts, deal structures and investments for Shariah compliance risks such as riba, gharar and exposure to prohibited activities, retrieves the relevant standards and fatwas, drafts the Shariah review documentation and flags issues for the Shariah board, which keeps sole authority over any ruling.",[154,197,125],[392,407,349],[176,132,29,110],[425],"Zoya","An internal assistant that screens contracts for compliance with Shariah standards is not listed in Annex III: it assesses contracts, structures and securities, not the creditworthiness of natural persons (Annex III point 5(b)). If a customer facing version answers product questions, it must disclose that people are interacting with an AI system under Article 50(1). National Islamic finance regulators set their own Shariah governance expectations.",{"slug":428,"title":429,"shortTitle":430,"definition":431,"status":19,"industries":432,"functions":433,"patterns":434,"audience":133,"autonomy":31,"adoptionStage":177,"evidenceCount":301,"publicEvidenceCount":301,"organizations":435,"bestGrade":41,"headline":438,"lastVerified":52,"indexable":12,"euAiActTier":146,"euAiActBasis":440},"tax-questions-and-filing-assistant","AI assistant for tax questions and filing support","Tax questions and filing assistant","An AI assistant that answers taxpayers' questions about taxes, deadlines, refunds and payments, lets authenticated taxpayers check their status or set up a payment plan within set rules, and guides them through filing, while assessments, penalties and disputes stay with the tax authority's staff and systems.",[251],[253,128,408],[130,131,176,29],[436,437,413],"ClearTax","HM Revenue and Customs",{"kpi":142,"label":143,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":439,"qualifier":98,"claimant":51,"organization":437,"vendorReported":11},83.03,"A taxpayer assistant must tell people they are interacting with an AI system (Article 50). It is not listed in Annex III as long as it only informs and applies fixed rules. It becomes high risk under Annex III point 5(a) if it evaluates eligibility for, or grants, reduces, revokes or reclaims, public assistance benefits (which can include benefits paid through the tax system). Recital 59 says systems used for administrative proceedings by tax and customs authorities are not high risk law enforcement systems; audit selection and risk scoring are covered on a separate page.",{"slug":442,"title":443,"shortTitle":444,"definition":445,"status":19,"industries":446,"functions":448,"patterns":449,"audience":65,"autonomy":111,"adoptionStage":177,"evidenceCount":34,"publicEvidenceCount":34,"organizations":450,"bestGrade":41,"headline":456,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":459},"developer-coding-assistant","AI coding assistant for software developers","Developer coding assistant","An AI assistant in the developer's IDE and code review flow that completes and generates code, explains unfamiliar modules, drafts unit tests and reviews pull requests for common defects, while generated code goes through the same review, testing and change controls as any other code.",[61,154,447,62,334],"capital-markets",[24],[299],[451,452,181,453,454,455],"Accenture","ANZ","Citi","CME Group","Meta",{"kpi":95,"label":96,"unit":45,"n":161,"nUpTo":47,"kind":457,"value":458,"qualifier":98,"claimant":165,"organization":165,"vendorReported":11},"median",20,"A coding assistant used by developers is not a prohibited practice under Article 5 and is not listed in Annex III. Developers know they are working with an AI tool, so the Article 50 disclosure duty has no practical effect for the deploying organization, and the marking of generated content under Article 50(2) falls on the tool's provider. What remains is AI literacy (Article 4). Using an AI system to monitor or evaluate individual developers' performance would fall under Annex III point 4(b), and the software the assistant helps build may itself fall under the Act.",{"slug":461,"title":462,"shortTitle":463,"definition":464,"status":19,"industries":465,"functions":466,"patterns":467,"audience":65,"autonomy":90,"adoptionStage":66,"segment":468,"evidenceCount":67,"publicEvidenceCount":67,"organizations":469,"bestGrade":41,"headline":472,"lastVerified":76,"indexable":12,"euAiActTier":53,"euAiActBasis":474},"hospital-bed-and-staff-capacity-command-center","AI command center for hospital bed and staff capacity planning","Hospital capacity command center","An AI powered operations center that predicts patient admissions, discharges and transfers across a hospital or health system, and helps a team of coordinators sitting in one room sequence real time bed assignments, staffing levels and patient moves, so patients get into the right bed faster and existing capacity is used fully without adding beds.",[173],[158,87],[28,29,27],"hospital operations",[470,471],"Humber River Health","Johns Hopkins Medicine",{"kpi":43,"label":44,"unit":45,"n":67,"nUpTo":47,"kind":48,"value":473,"qualifier":98,"claimant":51,"organization":471,"vendorReported":11},38,"The tier depends on what the system is scoped to do. A design limited to occupancy and discharge forecasting and to sequencing bed assignments for patients already admitted is operational decision support for hospital logistics, outside Annex III. Annex III point 5(d) covers AI used \"to dispatch, or to establish priority in the dispatching of, emergency first response services\", including medical aid and emergency healthcare patient triage systems. On a plain reading, that point can apply when a system dispatches, or sets the priority of dispatching, ambulance or critical care transport itself (work similar to what the Johns Hopkins center's Lifeline transport staff do for helicopter and ambulance transfers), or when it assesses the clinical urgency of an emergency patient, that is, triage. Sequencing which already admitted ED patient gets the next ward bed, and deciding whether to accept an inter hospital transfer request on capacity grounds, are not listed activities under 5(d) as written; whether either counts as dispatching or triage in a given deployment is a case by case legal question, not a settled fact, and should be assessed with counsel before relying on this tier. For public hospitals, Annex III point 5(a) (access to essential public services, including healthcare) can also be relevant. Scoping the system to bed sequencing and transfer acceptance only, and keeping every ambulance dispatch and ED triage decision with clinical staff outside the AI's recommendation, is what keeps a deployment in the lower tier.",{"slug":476,"title":477,"shortTitle":478,"definition":479,"status":19,"industries":480,"functions":481,"patterns":482,"audience":65,"autonomy":111,"adoptionStage":66,"segment":300,"evidenceCount":161,"publicEvidenceCount":161,"organizations":483,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":486},"client-briefing-and-call-report-copilot","AI copilot for corporate client briefings and call reports","Client briefing and call reports","An AI copilot for relationship managers, mainly in corporate and commercial banking, whose main job is preparation: before a client meeting it assembles a briefing pack from filings, news, internal notes, product holdings and upcoming maturities, and afterwards it turns the banker's notes into a structured call report and CRM update. Unlike a meeting notetaker, which centres on capturing the conversation, it centres on the credit and cross sell context around the meeting; wealth advisor tools that also prepare meetings overlap with it. The banker reviews every output.",[154,197,447],[319,233],[176,89,110,26],[181,484,485],"Scotiabank","Standard Chartered","Bankers interact with the copilot directly, but Article 50(1) does not bite here: it requires telling people they are dealing with an AI system unless that is obvious to a reasonably well informed person, and an internal tool that is openly presented and labelled as an AI assistant meets that bar by design. The copilot never interacts with the client. Article 50(2) marking of generated text falls on the provider of the system, including a bank that builds it in house, but the copilot turns a banker's own notes into a call report, an assistive function for standard editing of the banker's input that does not substantially alter it, so the Article 50(2) exception applies and no machine readable marking is required. It is not an Annex III use: credit context about corporate clients is not the creditworthiness assessment of natural persons in Annex III point 5(b), so it falls outside the high risk tier. If a deployment starts to score individuals for credit, the tier changes. AI literacy duties under Article 4 still apply. If meeting capture is used, recording and transcription rules under data protection law apply separately.",{"slug":488,"title":489,"shortTitle":490,"definition":491,"status":19,"industries":492,"functions":493,"patterns":494,"audience":65,"autonomy":111,"adoptionStage":66,"segment":495,"evidenceCount":134,"publicEvidenceCount":134,"organizations":496,"bestGrade":41,"headline":502,"lastVerified":52,"indexable":12,"euAiActTier":53,"euAiActBasis":503},"insurance-pricing-and-actuarial-copilot","AI copilot for insurance pricing and actuarial analysis","Pricing and actuarial copilot","AI that speeds up the work of pricing and actuarial teams, from automated, transparent risk and demand model building to natural language analysis of rate filings, experience data and reserving diagnostics, while actuaries select the models, sign off the rates and own the professional judgment.",[125],[349,393,87],[28,299,26,89],"pricing",[497,498,499,500,501],"Accelerant Holdings","Europ Assistance","Generali France","Kinsale Capital Group","MAIF",{"kpi":95,"label":96,"unit":326,"n":46,"nUpTo":47,"kind":48,"value":134,"qualifier":98,"claimant":51,"organization":499,"vendorReported":11},"Pricing and risk assessment of natural persons for life and health insurance is high risk under Annex III point 5(c). Pricing for property and casualty products, and actuarial analysis that does not price individuals, are not listed, although supervisors still expect sound model governance.",{"slug":505,"title":506,"shortTitle":507,"definition":508,"status":19,"industries":509,"functions":511,"patterns":513,"audience":65,"autonomy":111,"adoptionStage":66,"evidenceCount":134,"publicEvidenceCount":161,"organizations":514,"bestGrade":41,"headline":518,"lastVerified":145,"indexable":12,"euAiActTier":146,"euAiActBasis":520},"marketing-content-compliance-copilot","AI copilot for marketing content with compliance pre review","Marketing content and compliance","A copilot that drafts campaign copy, product explainers and social posts on brand and in the customer's language from approved product facts, then runs a first pass compliance check against advertising rules and required disclosures, flagging unsupported claims and missing warnings before a human in marketing compliance approves publication.",[61,154,125,155,197,510],"pharma-and-life-sciences",[512,392,407],"marketing",[110,176,29,216],[515,516,517],"Ally Financial","JPMorgan Chase","Klarna",{"kpi":95,"label":96,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":519,"qualifier":98,"claimant":51,"organization":515,"vendorReported":11},34,"An internal drafting and review aid that makes no decisions about people. Article 50 transparency duties apply to generated content: providers must mark synthetic content, and deployers must disclose deep fake images, audio or video. Personalized targeting of individuals is governed mainly by data protection and consumer law rather than the AI Act.",{"slug":522,"title":523,"shortTitle":524,"definition":525,"status":19,"industries":526,"functions":527,"patterns":528,"audience":65,"autonomy":111,"adoptionStage":32,"segment":529,"evidenceCount":161,"publicEvidenceCount":161,"organizations":530,"bestGrade":41,"headline":165,"lastVerified":76,"indexable":12,"euAiActTier":77,"euAiActBasis":533},"model-risk-validation-copilot","AI copilot for model risk validation and monitoring","Model risk validation","A copilot for independent model validation and review, whether run by a bank's validation function, an external tester or a supervisor, that checks model documentation against the model risk standard, generates and scores challenger tests (for generative AI, often with an LLM as a judge calibrated against human experts), watches production models for drift and drafts and consistency checks the validation report. An accountable validator owns every conclusion.",[154,125,447,197],[393,392],[26,132,110,27],"second-line",[531,485,532],"European Central Bank (ECB Banking Supervision)","United Overseas Bank (UOB)","A validation copilot supports internal governance and is not itself an Annex III use, and its drafts are internal, so Article 50 transparency duties do not normally apply. It often helps validate models that are high risk under Annex III (point 5(b), creditworthiness and credit scoring of natural persons; point 5(c), life and health insurance pricing), and the testing and documentation it supports feed the provider obligations of Articles 9, 11 and 15.",{"slug":535,"title":536,"shortTitle":537,"definition":538,"status":19,"industries":539,"functions":540,"patterns":541,"audience":65,"autonomy":111,"adoptionStage":66,"segment":33,"evidenceCount":34,"publicEvidenceCount":34,"organizations":542,"bestGrade":41,"headline":546,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":547},"network-fault-triage-copilot","AI copilot for network operations centre fault triage","NOC fault triage copilot","AI in the network operations centre (NOC) that correlates alarms and performance data from radio, transport, core and fixed networks into a small number of probable faults, ranks them by customer impact, proposes the likely root cause and fix from runbooks, vendor documentation and past tickets, and routes the ticket to the right team, while an engineer decides what to change.",[21],[23,158],[27,29,176,89,26],[543,36,38,544,40,545],"Bell Canada","Orange","Vodafone",{"kpi":43,"label":44,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":49,"qualifier":50,"claimant":51,"organization":36,"vendorReported":11},"The main test is Annex III point 2, which lists AI systems intended as safety components in the management and operation of critical digital infrastructure as high risk. A copilot that prepares diagnoses for engineers who decide every change is normally not such a safety component, and is then minimal risk. The tier rises when the system is designed to protect the safe operation of the network, for example by acting on it automatically to prevent or contain outages. Article 6(3) can exempt an Annex III system that only performs a preparatory task to an assessment and poses no significant risk of harm, provided the provider documents that assessment and registers the system.",{"slug":549,"title":550,"shortTitle":551,"definition":552,"status":19,"industries":553,"functions":555,"patterns":556,"audience":65,"autonomy":90,"adoptionStage":66,"segment":557,"evidenceCount":161,"publicEvidenceCount":161,"organizations":558,"bestGrade":41,"headline":562,"lastVerified":145,"indexable":12,"euAiActTier":146,"euAiActBasis":567},"plant-operator-and-maintenance-copilot","AI copilot for plant operators and maintenance technicians","Plant operator and maintenance copilot","A generative AI assistant for the people who run and repair machines in plants, workshops and service centres: it answers fault and procedure questions from equipment manuals, fault reports, shift logs and live machine data, in the technician's language, with links to the sources, so faults are diagnosed faster and expert knowledge is not lost when experienced staff retire.",[404,554],"automotive",[158,233],[176,130,89,216],"production",[559,560,561],"BMW Group","Georgia-Pacific","Textron Aviation",{"kpi":563,"label":564,"unit":565,"n":47,"nUpTo":46,"kind":48,"value":566,"qualifier":224,"claimant":99,"organization":561,"vendorReported":12},"time-saved-per-task","Time saved per task","minutes",18,"Article 50(1): staff must know they are interacting with an AI system, unless that is obvious from the context. Answering maintenance questions is not an Annex III use. It would become high risk under Annex III point 4(b) if the usage data were used to monitor and evaluate the performance of individual workers, so keep usage analytics aggregated.",{"slug":569,"title":570,"shortTitle":571,"definition":572,"status":19,"industries":573,"functions":574,"patterns":575,"audience":65,"autonomy":111,"adoptionStage":32,"segment":160,"evidenceCount":301,"publicEvidenceCount":301,"organizations":576,"bestGrade":41,"headline":165,"lastVerified":52,"indexable":12,"euAiActTier":77,"euAiActBasis":581},"suspicious-activity-report-drafting","AI copilot for SAR and STR narrative drafting","SAR and STR drafting","Generative AI that drafts the narrative of a single suspicious activity or suspicious transaction report from the investigation file (who, what, when, where, why and how), with every fact linked to its source record, so the investigator verifies, edits and files instead of starting from a blank page. It works case by case, unlike the periodic data returns of regulatory reporting.",[154,155],[200,254],[110,89,176,26],[577,578,579,580],"Finshark","BMO and Amalgamated Bank","Nexo","Uphold","Drafting internal reports for a human investigator is not listed in Annex III (the law enforcement uses in point 6 cover systems used by or for law enforcement authorities, not a bank's own reporting), and the text is not published to inform the public, so the deployer disclosure duty for generated text in Article 50(4) does not apply. Confidentiality rules for suspicious activity reports and GDPR apply in full.",{"slug":583,"title":584,"shortTitle":585,"definition":586,"status":19,"industries":587,"functions":588,"patterns":590,"audience":65,"autonomy":111,"adoptionStage":66,"segment":589,"evidenceCount":178,"publicEvidenceCount":178,"organizations":591,"bestGrade":41,"headline":599,"lastVerified":52,"indexable":12,"euAiActTier":53,"euAiActBasis":600},"underwriting-risk-assessment-copilot","AI copilot for underwriting risk assessment","Underwriting risk assessment copilot","A copilot that assembles everything relevant to a risk (the submission, loss history, internal guidelines, third party data and public information), highlights exposures and gaps against the insurer's underwriting guidelines and drafts the underwriting narrative or referral note, while the underwriter makes and signs every decision.",[125],[589,393],"underwriting",[176,89,110,26],[497,592,593,594,595,596,597,598],"American International Group","Arch Capital Group","Bowhead Specialty","Generali Global Corporate & Commercial","Hiscox","Skyward Specialty Insurance Group","Zurich North America",{"kpi":43,"label":44,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":97,"qualifier":98,"claimant":99,"organization":595,"vendorReported":12},"For commercial property and casualty lines the copilot is not listed in Annex III. Used for risk assessment of natural persons in life or health insurance it falls under Annex III point 5(c) and is high risk, with risk management, data governance, logging and human oversight duties, and deployers must carry out a fundamental rights impact assessment under Article 27.",{"slug":602,"title":603,"shortTitle":604,"definition":605,"status":19,"industries":606,"functions":607,"patterns":608,"audience":30,"autonomy":31,"adoptionStage":177,"evidenceCount":256,"publicEvidenceCount":134,"organizations":609,"bestGrade":41,"headline":615,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":616},"intelligent-document-processing","AI document intelligence for unstructured forms and documents","Intelligent document processing","AI that takes documents in any format, such as scanned forms, PDFs, photos, emails and handwritten notes, splits and classifies them, extracts the required fields with a confidence score, validates them against business rules and source systems, and sends only the uncertain cases to a person before the data enters the downstream process.",[61,251,554,404],[158,254,408],[132,321,29],[610,611,612,613,614],"Ancine","Pupuk Indonesia","U.S. Immigration and Customs Enforcement","U.S. Citizenship and Immigration Services","Volvo Group",{"kpi":142,"label":143,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":144,"qualifier":50,"claimant":99,"organization":610,"vendorReported":12},"Classifying documents and extracting data for a person or process to use is usually minimal risk. Even inside an Annex III area, a system that only performs a narrow procedural task, such as splitting and classifying documents, can fall outside the high risk category under Article 6(3); the provider must document that assessment and register the system (Article 6(4) and Article 49(2)). The picture changes when extraction materially influences decisions in Annex III areas, such as eligibility for public assistance benefits (point 5(a)), creditworthiness (point 5(b)) or asylum, visa and residence permit applications (point 7), where the whole system must be assessed as potentially high risk. The Article 6(3) exception never applies when the system performs profiling of natural persons.",{"slug":618,"title":619,"shortTitle":620,"definition":621,"status":19,"industries":622,"functions":623,"patterns":624,"audience":65,"autonomy":111,"adoptionStage":32,"evidenceCount":134,"publicEvidenceCount":134,"organizations":625,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":630},"civil-servant-drafting-copilot","AI drafting copilot for civil servants for correspondence, briefings and ministerial replies","Civil servant drafting copilot","A generative AI assistant that drafts replies to correspondence from the public and elected representatives, briefings, submissions and summaries for civil servants, grounded in the department's approved lines, policy documents and case data, with the official editing and approving every word before it is sent or cleared.",[251],[253,233,254],[110,176,89],[626,627,628,629,284],"Cabinet Office (Government Communication Service)","Crown Prosecution Service","Department for Education","Department for Science, Innovation and Technology (Incubator for Artificial Intelligence)","An internal drafting assistant that an official reviews is not listed in Annex III. Article 50(4) requires disclosure of AI generated text published to inform the public on matters of public interest, unless it has undergone human review and a person holds editorial responsibility, which this design provides. If the tool is used to evaluate eligibility for public assistance benefits or services rather than to draft, Annex III point 5(a) can apply.",{"slug":632,"title":633,"shortTitle":634,"definition":635,"status":19,"industries":636,"functions":637,"patterns":638,"audience":65,"autonomy":111,"adoptionStage":66,"evidenceCount":67,"publicEvidenceCount":67,"organizations":639,"bestGrade":41,"headline":642,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":645},"clinical-and-regulatory-document-drafting","AI drafting of clinical study reports and regulatory documents","Clinical and regulatory document drafting","Generative AI that drafts clinical study reports and other regulated documents, such as protocols, patient materials and submission modules, from the trial's statistical tables, listings and figures and from approved template text, for medical writers to verify, edit and approve before anything is submitted to a regulator.",[510],[392,158],[110,176,132],[640,641],"Merck & Co.","Novo Nordisk",{"kpi":643,"label":644,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":97,"qualifier":98,"claimant":51,"organization":640,"vendorReported":11},"error-reduction","Error reduction","Drafting regulated documents for expert review is not listed in Annex III and is not a practice prohibited by Article 5, so the tier turns on the sponsor's role under Article 50. A sponsor that deploys a third party drafting tool has no specific AI Act obligations beyond AI literacy: the Article 50(4) disclosure duty covers AI generated text published to inform the public on matters of public interest, which clinical study reports and regulatory submissions are not. For that sponsor the tier is minimal. A sponsor that builds its own generating system, as Merck (a proprietary platform) and Novo Nordisk (NovoScribe) did, is its provider under Article 50(2) and must mark the synthetic text in a machine readable format, unless the exemption for an assistive function for standard editing applies; drafting whole report sections goes beyond that exemption, so for that sponsor the tier is limited. Quality expectations come from medicines regulation and EMA guidance: the EMA reflection paper expects close human supervision and quality review when AI drafts medicinal product information documents, and makes the clinical trial sponsor, marketing authorisation applicant or holder, or manufacturer responsible for ensuring that models and data pipelines are fit for purpose and meet GxP standards and EMA guidelines.",{"slug":647,"title":648,"shortTitle":649,"definition":650,"status":19,"industries":651,"functions":652,"patterns":653,"audience":65,"autonomy":90,"adoptionStage":177,"evidenceCount":301,"publicEvidenceCount":301,"organizations":654,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":146,"euAiActBasis":657},"enterprise-knowledge-search","AI enterprise knowledge search for employees","Enterprise knowledge search","An assistant that lets any employee ask a question in plain language and get a synthesized answer from the organization's own policies, procedures, product manuals and research, with citations to the source documents and only from documents the employee is allowed to see.",[61,154,197,125,251,334],[233,158,128],[176,130,89],[181,396,655,656],"SIGNAL IDUNA","Wells Fargo","Article 50(1) requires that people who interact directly with an AI system are informed of it, unless this is obvious from the context, as it usually is for an internal assistant. The system would be high risk only if it were intended for an Annex III purpose, such as assessing the creditworthiness of natural persons (point 5(b)) or making decisions on or evaluating workers (point 4(b)).",{"slug":659,"title":660,"shortTitle":661,"definition":662,"status":19,"industries":663,"functions":664,"patterns":665,"audience":30,"autonomy":31,"adoptionStage":66,"segment":300,"evidenceCount":161,"publicEvidenceCount":161,"organizations":666,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":670},"trade-document-examination","AI examination of trade documents under letters of credit and collections","Trade document examination","AI that reads the full document presentation under a letter of credit or collection (bill of lading, commercial invoice, packing list, certificates), extracts and cross checks the data, tests it against the instructions and the ICC rules (for letters of credit, the credit terms, UCP 600 and ISBP), and lists discrepancies by severity with the rule cited, so qualified examiners focus on the genuine exceptions.",[154],[158,392],[132,29,26,89],[667,668,669],"ANZ, HSBC and Lloyds Banking Group","Rand Merchant Bank","Stanbic Bank Uganda","Checking trade documents for compliance with credit terms is not listed in Annex III and does not decide about natural persons. AI literacy duties under Article 4 apply, and the process falls under the bank's operational resilience and model governance.",{"slug":672,"title":673,"shortTitle":674,"definition":675,"status":19,"industries":676,"functions":677,"patterns":678,"audience":133,"autonomy":31,"adoptionStage":66,"segment":202,"evidenceCount":178,"publicEvidenceCount":34,"organizations":679,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":685},"financial-wellbeing-coach","AI financial wellbeing coach in the banking app","Financial wellbeing coach","An in app AI assistant that the customer opens to understand their own money: it uses the customer's transaction data to explain their spending, forecast upcoming bills and cash flow, set and track savings goals and answer money questions in plain language, staying on the guidance side of the line between guidance and regulated financial advice.",[154],[128,512],[130,376,28,26],[181,680,681,682,683,684],"Commonwealth Bank of Australia","Hyundai Card","Royal Bank of Canada","Starling Bank","Westpac","The conversational assistant carries the Article 50 transparency duty: customers must be told they are interacting with an AI system. The system becomes high risk if it is used to evaluate the creditworthiness of natural persons or establish their credit score (Annex III point 5(b)). Article 5(1)(b) prohibits AI that exploits vulnerabilities due to a person's specific social or economic situation to materially distort their behaviour in a way that causes, or is reasonably likely to cause, significant harm.",{"slug":687,"title":688,"shortTitle":689,"definition":690,"status":19,"industries":691,"functions":692,"patterns":693,"audience":65,"autonomy":31,"adoptionStage":66,"segment":160,"evidenceCount":178,"publicEvidenceCount":178,"organizations":694,"bestGrade":41,"headline":699,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":703},"aml-alert-triage","AI for AML transaction monitoring alert triage","AML alert triage","Machine learning and AI agents that score anti money laundering alerts for genuine risk, close clear false positives with a written and stored rationale, and hand investigators the remaining alerts already enriched with the customer, counterparty and transaction context.",[154,155],[200],[28,27,26,89],[695,578,696,579,697,698,532,580],"Australia Post","HSBC","Ratepay","Shift4",{"kpi":700,"label":701,"unit":45,"n":67,"nUpTo":47,"kind":48,"value":702,"qualifier":98,"claimant":99,"organization":698,"vendorReported":12},"false-positive-reduction","False positive reduction",86,"AML transaction monitoring is not listed in Annex III; point 5(b) covers creditworthiness and credit scoring and excludes systems used to detect financial fraud. The Article 5(1)(d) ban on predicting criminal offences from profiling alone does not apply to systems that support a human assessment already based on objective and verifiable facts linked to criminal activity, which is how alert triage should be designed. A decision to restrict an account taken solely by automated means would fall under GDPR Article 22 and national AML law, so consequential decisions need human review.",{"slug":705,"title":706,"shortTitle":707,"definition":708,"status":19,"industries":709,"functions":710,"patterns":711,"audience":30,"autonomy":90,"adoptionStage":66,"evidenceCount":134,"publicEvidenceCount":134,"organizations":712,"bestGrade":41,"headline":717,"lastVerified":145,"indexable":12,"euAiActTier":117,"euAiActBasis":721},"benefit-fraud-and-error-detection","AI for benefit fraud and error detection in social security","Benefit fraud and error detection","Risk models that help a social security or benefits agency decide which claims, payments and recipients to check for fraud or error, so that caseworkers verify the riskiest cases first, while every decision on entitlement stays with a person and the model is tested for fairness before and during use.",[251],[157,253,254],[28,27],[713,258,714,715,716],"Centers for Medicare and Medicaid Services","Gemeente Rotterdam","U.S. Department of the Treasury, Bureau of the Fiscal Service","Uitvoeringsinstituut Werknemersverzekeringen (UWV)",{"kpi":718,"label":719,"unit":326,"n":46,"nUpTo":47,"kind":48,"value":720,"qualifier":98,"claimant":51,"organization":258,"vendorReported":11},"detection-rate-improvement","Detection improvement",2.5,"Annex III point 5(a): AI systems used by or on behalf of public authorities to evaluate the eligibility of natural persons for essential public assistance benefits and services, or to grant, reduce, revoke or reclaim them. A fundamental rights impact assessment (Article 27) is required before a public body deploys it. A design that scores people over time on their social behaviour or personal characteristics and leads to unrelated or disproportionate detrimental treatment would fall under the Article 5(1)(c) prohibition on social scoring.",{"slug":723,"title":724,"shortTitle":725,"definition":726,"status":19,"industries":727,"functions":728,"patterns":729,"audience":30,"autonomy":31,"adoptionStage":32,"segment":300,"evidenceCount":161,"publicEvidenceCount":161,"organizations":730,"bestGrade":71,"headline":733,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":735},"business-onboarding-and-ubo-discovery","AI for business onboarding (KYB) and beneficial ownership discovery","Business onboarding and UBO","An AI agent that builds the know your business (KYB) due diligence file for a new or reviewed corporate client, before any account is opened: it collects registry, incorporation and ownership documents, resolves the entity across sources, maps the ownership chain through holding companies, nominees and trusts to the ultimate beneficial owners, screens the entity and its owners, and presents a risk scored case for a compliance analyst to decide.",[154,155,447],[199,200],[132,26,29,89],[731,732,324],"BNY","Incore Bank",{"kpi":221,"label":222,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":734,"qualifier":98,"claimant":51,"organization":731,"vendorReported":11},25,"Customer due diligence on legal entities is not listed in Annex III, and an internal analyst tool usually carries no Article 50 transparency duty, so the system is usually minimal risk. The design decides the rest: biometric verification that only confirms a director is who they claim to be is excluded from Annex III point 1(a), but remote biometric identification (one to many matching) is high risk, and so is any use of the output to assess the creditworthiness of the natural persons involved (point 5(b)). GDPR applies to the personal data of owners and directors throughout. Keep biometric and credit steps in separately assessed components.",{"slug":737,"title":738,"shortTitle":739,"definition":740,"status":19,"industries":741,"functions":742,"patterns":743,"audience":30,"autonomy":31,"adoptionStage":66,"segment":127,"evidenceCount":277,"publicEvidenceCount":256,"organizations":744,"bestGrade":41,"headline":748,"lastVerified":52,"indexable":12,"euAiActTier":53,"euAiActBasis":751},"claims-triage-and-straight-through-processing","AI for claims triage and straight through processing","Claims triage and STP","AI that reads each new insurance claim and its documents, scores its complexity, cover questions, fraud and recovery signals, sends it to the right handling path and handler, and settles simple, low risk claims end to end within set limits without a person touching them.",[125],[127,158],[29,28,132,26,89],[745,218,596,138,746,747,140],"Admiral Seguros","Sedgwick","Tokio Marine & Nichido Fire Insurance",{"kpi":221,"label":222,"unit":45,"n":46,"nUpTo":46,"kind":48,"value":749,"qualifier":750,"claimant":51,"organization":138,"vendorReported":11},55,"approximately","Claims handling as such is not listed in Annex III. The same system becomes high risk when it is also used for risk assessment and pricing of natural persons in life and health insurance (point 5(c)), or when it is used by or on behalf of a public authority to grant, reduce, revoke or reclaim essential public assistance benefits and services, including healthcare services (point 5(a)). Otherwise the tier is minimal, so the design and the operator decide. Decisions on claims based solely on automated processing are also subject to Article 22 of the GDPR and the UK GDPR.",{"slug":753,"title":754,"shortTitle":755,"definition":756,"status":19,"industries":757,"functions":758,"patterns":759,"audience":30,"autonomy":31,"adoptionStage":66,"segment":589,"evidenceCount":277,"publicEvidenceCount":277,"organizations":760,"bestGrade":41,"headline":765,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":767},"commercial-underwriting-submission-triage","AI for commercial underwriting submission intake and triage","Underwriting submission triage","AI that reads incoming broker submissions for commercial insurance (emails, applications, schedules of values, loss runs and supplements), extracts the risk data into a structured record, checks clearance and appetite, enriches the risk with internal and third party data and ranks it, so underwriters open a complete, prioritized file instead of an inbox.",[125],[589,158],[132,29,28,26],[592,761,762,595,596,500,763,764,597],"AXIS Capital","CNA Financial","Markel","Paragon Insurance Group",{"kpi":142,"label":143,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":766,"qualifier":750,"claimant":51,"organization":764,"vendorReported":11},98,"Intake and triage for commercial insurance is not listed in Annex III, which covers risk assessment and pricing of natural persons in life and health insurance. It moves up to high risk only if the same pipeline is used to assess or price life or health cover for individuals.",{"slug":769,"title":770,"shortTitle":771,"definition":772,"status":19,"industries":773,"functions":774,"patterns":775,"audience":30,"autonomy":111,"adoptionStage":32,"segment":529,"evidenceCount":161,"publicEvidenceCount":161,"organizations":776,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":779},"complaints-root-cause-analysis","AI for complaints root cause and systemic issue analysis","Complaints root cause analysis","AI that reads the free text of complaints across all channels, clusters them into themes, separates systemic causes from one off events, links each theme to the product, process or control behind it and routes the insight to the owner who can fix it, with a human validating every root cause and every remediation.",[61,154,125,155,21,251],[392,128,87],[29,89,26,176],[713,777,778],"Board of Governors of the Federal Reserve System","Federal Trade Commission","Analysing complaints in aggregate to find causes is not listed in Annex III, is not a practice prohibited by Article 5 and does not decide on individuals. It does not interact with the public, so the disclosure duty in Article 50(1) does not apply; the machine readable marking of generated text in Article 50(2) is a duty of the provider of the generative model or system that writes the summaries. If the same system decided individual complaint outcomes or redress, or its themes were used to evaluate the performance of individual complaint handlers (Annex III point 4), that design would need its own assessment.",{"slug":781,"title":782,"shortTitle":783,"definition":784,"status":19,"industries":785,"functions":786,"patterns":787,"audience":30,"autonomy":31,"adoptionStage":32,"segment":529,"evidenceCount":161,"publicEvidenceCount":161,"organizations":788,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":792},"continuous-controls-testing","AI for continuous controls testing and control self assessment","Continuous controls testing","AI that moves control testing from periodic samples to continuous, full population assurance: it collects evidence from source systems, maps each artefact to the control it supports, tests every transaction or record against the control's rule, flags exceptions for a human to judge and prepares the risk and control self assessment from incident and loss data for the business to review.",[61,154,125,447,251],[393,392,158],[26,132,27,29],[789,790,791],"Federal Deposit Insurance Corporation","U.S. Department of the Interior","Pension Benefit Guaranty Corporation","Testing controls over transactions and systems is not an Annex III use. Controls that monitor and evaluate individual employees' behaviour, such as trading or access conduct, can fall under Annex III point 4(b), so the design decides the tier.",{"slug":794,"title":795,"shortTitle":796,"definition":797,"status":19,"industries":798,"functions":799,"patterns":800,"audience":65,"autonomy":111,"adoptionStage":66,"evidenceCount":301,"publicEvidenceCount":301,"organizations":801,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":805},"court-and-case-file-summarization","AI for court and case file summarization","Case file summarization","AI that condenses court filings, case files, evidence recordings and earlier decisions into structured summaries, chronologies and draft case reports with references to the source pages, so that judges, prosecutors, tribunal staff and government lawyers find what matters faster, while the person responsible reads the underlying material and makes every legal judgment.",[251],[407,254],[89,132,176],[627,802,803,804],"U.S. Department of Justice","Gemeente Amsterdam","Supremo Tribunal Federal","Annex III point 8(a) makes AI high risk when it is intended to assist a judicial authority in researching and interpreting facts and the law and in applying the law to a concrete set of facts. Tools for prosecutors fall under point 6(c) if they evaluate the reliability of evidence, and tools that assist the examination of asylum, visa or residence applications fall under point 7(c). Under Article 6(3) a system that only performs a narrow procedural task or a preparatory task, such as organising a file or transcribing and summarising it for the person who decides, may not be high risk, but the provider must document that assessment (Article 6(4)). Summaries of internal legal advice for government lawyers, as Amsterdam plans, are generally outside Annex III.",{"slug":807,"title":808,"shortTitle":809,"definition":810,"status":19,"industries":811,"functions":812,"patterns":813,"audience":65,"autonomy":111,"adoptionStage":66,"evidenceCount":134,"publicEvidenceCount":134,"organizations":814,"bestGrade":41,"headline":819,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":820},"training-content-generation","AI for creating employee training and eLearning content","Training content creation","Generative AI that helps learning and development teams turn source material such as procedures, product documentation and policies into training: course outlines, lesson text, quizzes, narration, avatar videos and translations, which instructional designers and subject matter experts review before publishing.",[61,251,62,404],[108,233],[110,216,89],[815,413,816,817,818],"Carlsberg Group","U.S. Marshals Service","Veterans Benefits Administration","Zoom",{"kpi":43,"label":44,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":144,"qualifier":98,"claimant":99,"organization":818,"vendorReported":12},"Generating training content is not listed in Annex III. Providers of tools that generate synthetic audio, image, video or text content must mark the output as AI generated (with an exception for assistive editing that does not substantially alter the source), and deployers must disclose deep fakes, such as an avatar or voice that resembles a real person and would falsely appear authentic (Article 50(2) and (4), with the definition in Article 3(60)). If the same system evaluates learning outcomes or decides access to training that affects a person's work, Annex III point 3 (education and vocational training) and point 4 (employment) must be checked, and those parts can be high risk.",{"slug":822,"title":823,"shortTitle":824,"definition":825,"status":19,"industries":826,"functions":827,"patterns":829,"audience":65,"autonomy":111,"adoptionStage":66,"segment":30,"evidenceCount":134,"publicEvidenceCount":134,"organizations":830,"bestGrade":41,"headline":834,"lastVerified":52,"indexable":12,"euAiActTier":146,"euAiActBasis":836},"outbound-notice-drafting","AI for drafting customer letters and outbound notices","Outbound notice drafting","AI that drafts the letters and notices operations must send at scale, such as arrears notices, decline letters, complaint responses, servicing confirmations and product change notices, from case data and approved templates and clauses, in the customer's language, for a person to approve where the notice is regulated.",[61,154,125,251,173,197],[158,128,828,392,127],"collections-and-recovery",[110,176,216],[831,596,832,833],"Acentra Health","Health Resources and Services Administration","SS&C Technologies",{"kpi":43,"label":44,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":734,"qualifier":98,"claimant":99,"organization":835,"vendorReported":12},"SS&C GIDS and RS","Drafting letters for human approval is not listed in Annex III. The decision the letter communicates may come from a separate high risk system, such as credit scoring (Annex III point 5(b)) or a public body's eligibility decision on benefits (point 5(a)); the drafting tool does not make that decision. Article 50(2) requires the provider of an AI system that generates text to mark the output as artificially generated, which puts this on the limited risk (transparency) tier; this includes an organization that builds its own drafting tool. Article 50(2) does not apply where the AI has only an assistive function for standard editing and does not substantially alter the input data or the semantics of the output.",{"slug":838,"title":839,"shortTitle":840,"definition":841,"status":19,"industries":842,"functions":843,"patterns":844,"audience":65,"autonomy":111,"adoptionStage":177,"evidenceCount":301,"publicEvidenceCount":301,"organizations":845,"bestGrade":41,"headline":848,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":850},"ediscovery-and-disclosure-document-review","AI for eDiscovery and disclosure document review","eDiscovery document review","AI that sorts, prioritises and codes large collections of emails, chats and files for relevance, issues and legal privilege in litigation, investigations and regulatory requests, so that lawyers review the documents most likely to matter and can show the court how the rest were handled.",[61,334,251],[407,254],[29,132,89],[802,778,846,847],"Purpose Legal","Serious Fraud Office",{"kpi":43,"label":44,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":849,"qualifier":98,"claimant":99,"organization":846,"vendorReported":12},85,"Document review for a party in civil litigation or an internal investigation is not listed in Annex III, so it is usually minimal risk. It becomes high risk where a law enforcement authority uses AI to evaluate the reliability of evidence in the investigation or prosecution of criminal offences (Annex III point 6(c)), or where a judicial authority uses it to research and interpret facts and law (point 8(a)). Prosecutors and investigators should classify each use against those points.",{"slug":852,"title":853,"shortTitle":854,"definition":855,"status":19,"industries":856,"functions":857,"patterns":858,"audience":65,"autonomy":111,"adoptionStage":66,"evidenceCount":301,"publicEvidenceCount":301,"organizations":859,"bestGrade":41,"headline":165,"lastVerified":52,"indexable":12,"euAiActTier":77,"euAiActBasis":862},"freedom-of-information-request-processing","AI for freedom of information request processing","Freedom of information requests","AI that helps a public body handle freedom of information and open government requests: logging and clarifying requests, spotting duplicates, searching and deduplicating the records in scope, proposing redactions with the exemption that applies, and drafting the response letter, with an FOI officer deciding what is released.",[251],[253,407,254],[132,29,110],[802,860,861,790],"U.S. Food and Drug Administration, Center for Drug Evaluation and Research","Provincie Noord-Holland","Tools that support staff in searching, deduplicating and proposing redactions are not listed in Annex III (point 5(a) covers eligibility for public assistance benefits and services, not access to documents), and every release decision stays with an officer. A public facing request assistant that talks to requesters would carry the Article 50(1) transparency duty.",{"slug":864,"title":865,"shortTitle":866,"definition":867,"status":19,"industries":868,"functions":869,"patterns":870,"audience":65,"autonomy":111,"adoptionStage":66,"segment":127,"evidenceCount":134,"publicEvidenceCount":134,"organizations":871,"bestGrade":41,"headline":874,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":875},"health-prior-authorization-and-claims-adjudication","AI for health insurance prior authorization and claims adjudication support","Health prior authorization and adjudication","AI that reads prior authorization requests, medical claims and appeals with their clinical and billing documents, extracts diagnoses, treatments and costs, checks them against the policy and published clinical criteria, and prepares a summary and recommendation for a clinician or adjudicator, who makes every adverse decision.",[125,173],[127,254,158],[132,89,176,29,110],[831,872,713,873,379],"AdvanceCare","ICICI Lombard",{"kpi":241,"label":242,"unit":45,"n":67,"nUpTo":47,"kind":48,"value":97,"qualifier":750,"claimant":99,"organization":831,"vendorReported":12},"Annex III point 5(a) makes AI high risk when it is used by or on behalf of public authorities to evaluate eligibility for essential public assistance benefits and services, including healthcare services, or to grant, reduce or revoke them, which can cover statutory health schemes run by or for public bodies. Point 5(c) covers risk assessment and pricing in life and health insurance, not claim review. A copilot for a private insurer's claim review, where people decide, is usually outside Annex III; for public schemes, Article 6(3) may exempt a system that only performs a preparatory task, unless it profiles natural persons. GDPR rules on health data (Article 9) and on solely automated decisions (Article 22) apply in every case.",{"slug":877,"title":878,"shortTitle":879,"definition":880,"status":19,"industries":881,"functions":882,"patterns":883,"audience":133,"autonomy":111,"adoptionStage":66,"evidenceCount":301,"publicEvidenceCount":301,"organizations":884,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":887},"immigration-and-visa-application-assistant","AI for immigration and visa applications, from applicant questions to case preparation","Immigration and visa application assistant","AI that helps applicants understand immigration and visa requirements and submit complete applications, and helps immigration staff prepare cases by extracting form data, classifying evidence, routing applications and supporting interviews, while every grant or refusal is decided by an officer against the immigration rules.",[251],[253,254,158],[130,132,29,216],[885,886,612,613],"Home Office (Visa, Status and Information Services)","U.S. Department of State (Bureau of Consular Affairs)","Annex III point 7(c) makes AI high risk when it assists public authorities in examining applications for asylum, visas or residence permits, including assessing the reliability of evidence. Applicant facing information assistants that give general guidance fall under the Article 50 transparency duties (limited risk). Evidence classification, routing and interview support used in the examination are likely high risk, unless the provider documents under Article 6(3) that a component only performs a narrow procedural or preparatory task. That exception never applies to a system that profiles natural persons, which matters for routing on personal attributes or risk profiles.",{"slug":889,"title":890,"shortTitle":891,"definition":892,"status":19,"industries":893,"functions":894,"patterns":895,"audience":30,"autonomy":90,"adoptionStage":177,"segment":127,"evidenceCount":134,"publicEvidenceCount":134,"organizations":896,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":899},"claims-fraud-detection","AI for insurance claims fraud detection","Claims fraud detection","AI that scores every insurance claim for fraud from first notice of loss onwards, combining claim, policy, document, image and network data to find suspicious claims, organised rings and inflated losses, and sends each alert with its reasons to a claims handler or special investigations unit for review.",[125],[127,157],[27,28,132,321,29],[897,898,219,138,747],"Assurant","AXA Switzerland","Claims fraud detection by an insurer is not listed in Annex III, and point 5(b) explicitly excludes AI systems used to detect financial fraud from the credit scoring category. Point 5(c) covers only risk assessment and pricing in life and health insurance, so a fraud model becomes high risk when it also feeds those decisions, or when it is used by or on behalf of a public authority to grant, reduce, revoke or reclaim public assistance benefits (point 5(a)). Profiling and automated decisions remain subject to GDPR, including Article 22 where a claim is refused on a decision based solely on automated processing.",{"slug":901,"title":902,"shortTitle":903,"definition":904,"status":19,"industries":905,"functions":906,"patterns":907,"audience":65,"autonomy":111,"adoptionStage":66,"evidenceCount":34,"publicEvidenceCount":134,"organizations":908,"bestGrade":41,"headline":913,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":914},"aiops-incident-triage","AI for IT incident triage and root cause analysis (AIOps)","AIOps incident triage","AI that turns a flood of monitoring alerts into one probable incident, routes it to the right team, proposes likely root causes and remediation from runbooks and past incidents, and drafts the stakeholder updates and the post incident review, while an engineer authorizes every change.",[61,154,62,21,155],[24,158,393],[27,29,89,176,26],[909,455,910,911,912],"Google","Microsoft","Mizuho Financial Group","TD Bank",{"kpi":142,"label":143,"unit":45,"n":161,"nUpTo":47,"kind":457,"value":144,"qualifier":98,"claimant":165,"organization":165,"vendorReported":11},"An internal tool that supports engineers on IT incidents; it is not a use listed in Annex III and makes no decisions about people. Annex III point 2 covers AI used as a safety component in the management and operation of critical digital infrastructure, and recital 55 limits safety components to systems that directly protect the physical integrity of that infrastructure or the health and safety of persons and property. A triage copilot that proposes causes and fixes to engineers does not normally do that, but operators of critical digital infrastructure (cloud, data centers, telecom networks) should confirm this for their own design.",{"slug":916,"title":917,"shortTitle":918,"definition":919,"status":19,"industries":920,"functions":921,"patterns":922,"audience":30,"autonomy":31,"adoptionStage":66,"segment":30,"evidenceCount":301,"publicEvidenceCount":301,"organizations":923,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":928},"ledger-and-payment-reconciliation","AI for ledger and payment reconciliation","Ledger and payment reconciliation","AI that matches entries across nostro and vostro statements, card and scheme settlement files, the general ledger and suspense accounts, proposes matches and clearing journals, and routes only the genuine breaks to an operator with a plain language explanation.",[154,155,447,61,197,251],[408,158],[26,27,132],[924,925,926,927],"Comrade Trustee Services","Ginnie Mae","National Bank of Greece (Cyprus)","World Food Programme","Matching entries between internal financial records is not a use listed in Annex III and is not a practice prohibited by Article 5. Operators knowingly use an internal AI tool, so no Article 50(1) disclosure is needed. If a generative model drafts the explanations or journals, the provider of that system may have to mark its output as AI generated under Article 50(2). The AI literacy duty of Article 4 applies to the bank as deployer.",{"slug":930,"title":931,"shortTitle":932,"definition":933,"status":19,"industries":934,"functions":935,"patterns":936,"audience":65,"autonomy":111,"adoptionStage":66,"evidenceCount":34,"publicEvidenceCount":134,"organizations":937,"bestGrade":41,"headline":941,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":942},"legacy-code-modernization","AI for legacy code modernization","Legacy code modernization","AI that reads legacy code such as COBOL, PL/I or old Java, explains what each program does, maps its data flows and dependencies, drafts the equivalent modern code or specification, and generates the regression tests needed to prove the new system behaves like the old one.",[61,154,447,554,62],[24],[299,89,26],[938,939,909,396,940],"Airbnb","Amazon","Toyota Motor Europe",{"kpi":43,"label":44,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":97,"qualifier":750,"claimant":51,"organization":909,"vendorReported":11},"Tools that analyze, document and translate code are not prohibited practices under Article 5 and are not listed in Annex III, so no high risk obligations apply to the tooling. Engineers and analysts know they are working with an AI tool, including when they query the documentation through a chat assistant, so the Article 50 disclosure duty has no practical effect for the deploying organization. What remains is AI literacy for the staff who use it (Article 4). If the system being modernized is itself an AI system in an Annex III area (for example creditworthiness assessment, point 5(b)), its new version still has to meet the high risk requirements.",{"slug":944,"title":945,"shortTitle":946,"definition":947,"status":19,"industries":948,"functions":949,"patterns":950,"audience":65,"autonomy":111,"adoptionStage":66,"segment":529,"evidenceCount":134,"publicEvidenceCount":134,"organizations":951,"bestGrade":41,"headline":956,"lastVerified":52,"indexable":12,"euAiActTier":53,"euAiActBasis":958},"market-abuse-surveillance-triage","AI for market abuse surveillance alert triage","Market abuse surveillance","AI that helps surveillance analysts triage market abuse and conduct alerts, such as spoofing, layering, wash trades, ramping and insider dealing, by gathering the trade, order, news and communications context, explaining in plain language what triggered each alert and drafting the investigation narrative for the analyst to disposition.",[447,154,197],[392,200],[27,26,89,29],[952,205,953,954,955],"Commodity Futures Trading Commission","Japan Exchange Group","Nasdaq","U.S. Securities and Exchange Commission",{"kpi":241,"label":242,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":957,"qualifier":750,"claimant":51,"organization":954,"vendorReported":11},33,"Surveillance of orders and transactions as such is not listed in Annex III. Where the system monitors and evaluates the behaviour of the firm's own staff, in their communications or their trading, it can fall under Annex III point 4(b) (AI used to monitor and evaluate the performance and behaviour of persons in work relationships), so the tier depends on whether the system scores individual employees. Inferring employees' emotions from biometric data such as voice recordings is prohibited in the workplace under Article 5(1)(f).",{"slug":960,"title":961,"shortTitle":962,"definition":963,"status":19,"industries":964,"functions":965,"patterns":966,"audience":65,"autonomy":31,"adoptionStage":66,"segment":33,"evidenceCount":134,"publicEvidenceCount":134,"organizations":967,"bestGrade":41,"headline":970,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":971},"network-planning-and-capacity-optimization","AI for mobile network planning and capacity optimization","Network planning and capacity","Machine learning that forecasts where and when a mobile network will run out of capacity, recommends where to add cells, spectrum or hardware, and continuously tunes radio parameters so existing capacity carries more traffic, with planners approving investments and major changes.",[21],[23,87],[28,376,27,26],[36,968,39,969,545],"NTT DOCOMO","Telefónica España",{"kpi":43,"label":44,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":49,"qualifier":50,"claimant":51,"organization":36,"vendorReported":11},"Forecasting demand, ranking congested cells and recommending investments is normally minimal risk. Under Article 6(2), Annex III point 2 lists AI systems intended as safety components in the management and operation of critical digital infrastructure as high risk, and Recital 55 ties this to the digital infrastructure in the Annex to Directive (EU) 2022/2557, which includes providers of public electronic communications networks. Recital 55 defines such safety components as systems that directly protect the physical integrity of the infrastructure or the health and safety of persons and property and that are not necessary for the system to function. Closed loop parameter optimisation on the live radio network is high risk only when it serves in that role, for example a loop whose purpose is to protect emergency call availability, so each automated loop should be assessed against point 2 and the outcome documented. Loops that only optimise performance or capacity are usually not safety components.",{"slug":973,"title":974,"shortTitle":975,"definition":976,"status":19,"industries":977,"functions":978,"patterns":979,"audience":133,"autonomy":31,"adoptionStage":66,"evidenceCount":256,"publicEvidenceCount":256,"organizations":980,"bestGrade":41,"headline":986,"lastVerified":52,"indexable":12,"euAiActTier":146,"euAiActBasis":988},"non-emergency-service-request-routing","AI for non emergency service requests and 311 routing","Non emergency service request routing","An AI agent on a city's 311 style phone, chat and messaging channels that answers routine municipal questions, takes service requests such as potholes, missed collections or broken street lights with the right location and details, creates the case in the work order system and routes anything urgent or complex to the right team.",[251],[253,128,254],[130,131,29,26],[279,981,982,983,287,984,985],"London Borough of Barnet","City of Kelowna","Galt Police Department","Newcastle City Council","Rio de Janeiro City Data Office (Escritório de Dados)",{"kpi":142,"label":143,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":987,"qualifier":98,"claimant":99,"organization":982,"vendorReported":12},80,"A 311 assistant must disclose that it is AI (Article 50). It is not high risk while it only informs and creates service cases. If it evaluates or classifies emergency calls or sets dispatch priority for police, fire or medical services, it falls under Annex III point 5(d) and becomes high risk.",{"slug":990,"title":991,"shortTitle":992,"definition":993,"status":19,"industries":994,"functions":995,"patterns":996,"audience":65,"autonomy":111,"adoptionStage":32,"evidenceCount":134,"publicEvidenceCount":134,"organizations":997,"bestGrade":41,"headline":1001,"lastVerified":52,"indexable":12,"euAiActTier":53,"euAiActBasis":1002},"permit-and-licence-application-processing","AI for permit and licence application processing","Permit and licence application processing","AI that helps applicants submit complete permit and licence applications and helps officers process them, by answering questions about requirements, checking applications for missing or inconsistent information, pulling the relevant policies, history and constraints, and drafting reports, while the grant or refusal stays with a named officer or a published rule.",[251],[253,254,392],[132,26,176,130],[998,262,999,340,1000],"Intellectual Property Office","U.S. Fish and Wildlife Service","West Berkshire Council",{"kpi":142,"label":143,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":849,"qualifier":50,"claimant":51,"organization":262,"vendorReported":11},"Permit and licence decisions are not listed as such in Annex III, so officer decision support is usually minimal risk, and an assistant that talks to applicants carries the Article 50 transparency duty. The exceptions are permits in an Annex III area: examining applications for visas and residence permits (point 7) and evaluating eligibility for essential public assistance benefits and services (point 5(a)) are high risk. Solely automated decisions with legal or similarly significant effects on a person fall under GDPR Article 22 whatever the tier.",{"slug":1004,"title":1005,"shortTitle":1006,"definition":1007,"status":19,"industries":1008,"functions":1009,"patterns":1010,"audience":30,"autonomy":31,"adoptionStage":66,"evidenceCount":301,"publicEvidenceCount":301,"organizations":1011,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1015},"adverse-event-case-intake","AI for pharmacovigilance adverse event case intake","Adverse event case intake","AI that takes in adverse event reports about medicines, vaccines and devices from calls, emails, forms, literature and partner files, decides whether each is a valid case, flags seriousness, extracts and codes the case data into the safety database format, and routes it to drug safety professionals, who review medical content and regulatory reporting.",[510,251],[392,254,158],[132,29,130],[1012,860,1013,1014],"Bayer","U.S. Food and Drug Administration","Pfizer","Internal intake, extraction and coding for review by safety staff is not listed in Annex III and is usually minimal risk. A public facing reporting assistant must tell people they are talking to an AI under Article 50. The main obligations come from pharmacovigilance law and good pharmacovigilance practices, which require validated, inspectable processes, and from GDPR rules on health data.",{"slug":1017,"title":1018,"shortTitle":1019,"definition":1020,"status":19,"industries":1021,"functions":1022,"patterns":1023,"audience":133,"autonomy":31,"adoptionStage":66,"segment":127,"evidenceCount":134,"publicEvidenceCount":134,"organizations":1024,"bestGrade":71,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":146,"euAiActBasis":1028},"photo-based-damage-assessment","AI for photo based damage assessment in insurance claims","Photo damage assessment","Computer vision that assesses damage from photos or video of a vehicle or property taken by the policyholder, a repairer or an adjuster, identifies the damaged parts and the repair or replace decision, produces or checks the repair estimate, and flags total losses and inconsistencies for a person to review.",[125],[127],[321,28,26],[745,1025,1026,1027,747],"Covéa","Foyer","PZU","Assessing damage to vehicles or property for property and casualty claims is not listed in Annex III, which covers insurance only for risk assessment and pricing of natural persons in life and health insurance (point 5(c)). Article 50(1) transparency duties apply when the customer interacts directly with the AI, for example a guided photo journey that returns an AI estimate or offer, or a chat agent. A purely internal repairer estimate review with no customer interaction is minimal. A settlement or refusal decided solely by automated processing can fall under GDPR Article 22.",{"slug":1030,"title":1031,"shortTitle":1032,"definition":1033,"status":19,"industries":1034,"functions":1035,"patterns":1036,"audience":65,"autonomy":111,"adoptionStage":32,"segment":529,"evidenceCount":161,"publicEvidenceCount":161,"organizations":1037,"bestGrade":41,"headline":165,"lastVerified":52,"indexable":12,"euAiActTier":77,"euAiActBasis":1038},"policy-drafting-and-gap-analysis","AI for policy drafting and policy gap analysis","Policy drafting and gaps","An assistant that takes a new or changed obligation, finds every internal policy, standard and procedure it touches, flags clauses that now conflict or are silent, and drafts the updated wording in house style as a redline for the policy owner to approve.",[61,154,125,447,251],[392,407,233],[176,110,132,89],[789,412,832],"Drafting internal policy text for human approval is not an Annex III use and has no direct effect on individuals. The Article 4 AI literacy measures still apply to the staff who use it.",{"slug":1040,"title":1041,"shortTitle":1042,"definition":1043,"status":19,"industries":1044,"functions":1045,"patterns":1047,"audience":30,"autonomy":31,"adoptionStage":66,"segment":33,"evidenceCount":34,"publicEvidenceCount":34,"organizations":1048,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1050},"predictive-network-maintenance","AI for predictive network maintenance in telecom","Predictive network maintenance","Machine learning that spots the early signs of network failure, such as degrading cells, faulty customer equipment, ageing hardware or planned digging near fibre, and triggers a preventive fix, a remote reset or a targeted intervention before customers lose service.",[21],[23,1046,158],"field-service",[27,28,26],[38,544,969,40,1049,545],"Verizon","Scoring failure risk and planning maintenance is normally minimal risk. Annex III point 2 lists AI systems intended as safety components in the management and operation of critical digital infrastructure as high risk, and public electronic communications networks fall within that infrastructure. Recital 55 limits safety components to systems that directly protect the physical integrity of the infrastructure or the health and safety of persons and property, and excludes components used solely for cybersecurity. An operator whose automated actions meet that test must treat the system as high risk.",{"slug":1052,"title":1053,"shortTitle":1054,"definition":1055,"status":19,"industries":1056,"functions":1057,"patterns":1058,"audience":30,"autonomy":111,"adoptionStage":66,"evidenceCount":134,"publicEvidenceCount":134,"organizations":1059,"bestGrade":41,"headline":1063,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":1065},"public-consultation-response-analysis","AI for public consultation response analysis","Consultation response analysis","AI that reads every free text response to a public consultation or rulemaking comment period, proposes themes, maps each response to the themes that officials have validated, flags duplicates, campaign letters and responses that need special attention, and produces counts and summaries for the analysts who write the government's response.",[251],[253,87],[89,29,110],[1060,777,1061,629,1062],"Centers for Disease Control and Prevention","Department for Transport","U.S. Department of Transportation, Office of the Secretary",{"kpi":142,"label":143,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":1064,"qualifier":50,"claimant":51,"organization":1061,"vendorReported":11},92,"Organising and summarising consultation responses for analysts does not decide on individuals and is not listed in Annex III, so no high risk obligations apply. If AI generated text is published to inform the public on matters of public interest without human review and editorial responsibility, Article 50(4) requires disclosure.",{"slug":1067,"title":1068,"shortTitle":1069,"definition":1070,"status":19,"industries":1071,"functions":1072,"patterns":1073,"audience":30,"autonomy":1074,"adoptionStage":66,"segment":33,"evidenceCount":134,"publicEvidenceCount":134,"organizations":1075,"bestGrade":41,"headline":1081,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1084},"ran-energy-optimization","AI for radio access network energy optimization","RAN energy optimization","Machine learning that predicts traffic per cell and puts radio carriers, cells and hardware components into sleep modes when demand is low, then wakes them before users notice, so a mobile network uses less electricity without losing coverage or quality.",[21],[23],[28],"autonomous",[1076,1077,1078,1079,1080],"BT Group","Indosat Ooredoo Hutchison","O2 Telefónica Germany","Safaricom","Telefónica",{"kpi":1082,"label":1083,"unit":45,"n":47,"nUpTo":46,"kind":48,"value":178,"qualifier":224,"claimant":51,"organization":1080,"vendorReported":11},"energy-savings","Energy savings","Optimizing energy use is normally minimal risk. Under Article 6(2), Annex III point 2 lists AI systems intended as safety components in the management and operation of critical digital infrastructure as high risk, and public electronic communications networks fall under that infrastructure. Recital 55 limits safety components to systems that directly protect the infrastructure or the health and safety of persons, so an optimizer is not high risk by default, but a design in which it could affect emergency service availability should be assessed against point 2.",{"slug":1086,"title":1087,"shortTitle":1088,"definition":1089,"status":19,"industries":1090,"functions":1091,"patterns":1092,"audience":133,"autonomy":111,"adoptionStage":66,"evidenceCount":134,"publicEvidenceCount":134,"organizations":1093,"bestGrade":41,"headline":1098,"lastVerified":145,"indexable":12,"euAiActTier":117,"euAiActBasis":1099},"recruitment-screening-and-interview-scheduling","AI for recruitment screening and interview scheduling","Recruitment screening and scheduling","AI that answers candidates' questions, collects applications in conversation, schedules interviews and, where the organization chooses, assesses applications against the job requirements for a recruiter, who makes every selection decision. In the EU, the screening part is a high risk AI system under Annex III point 4 of the AI Act.",[61,251,213,334],[108],[130,29,28,26],[1094,1095,612,1096,1097],"Chipotle Mexican Grill","Gojob","Mastercard","Trace3",{"kpi":43,"label":44,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":144,"qualifier":750,"claimant":51,"organization":1096,"vendorReported":11},"Annex III point 4(a) lists AI systems intended to be used for the recruitment or selection of natural persons, in particular to place targeted job advertisements, to analyse and filter job applications and to evaluate candidates. Screening, ranking and scoring applications is therefore high risk. A component limited to a narrow procedural task, such as booking interview slots or answering process questions, can fall outside the high risk category under Article 6(3), but only if it does not materially influence the outcome and does not profile people, and that assessment must be documented (Article 6(4)). Deployers of the high risk part must follow the instructions for use, assign competent human oversight, keep logs, inform workers' representatives and inform candidates that a high risk system is used (Article 26). An organization that builds its own screening system becomes its provider, with conformity assessment duties. The chatbot part also carries the Article 50 disclosure duty.",{"slug":1101,"title":1102,"shortTitle":1103,"definition":1104,"status":19,"industries":1105,"functions":1106,"patterns":1107,"audience":65,"autonomy":111,"adoptionStage":66,"evidenceCount":301,"publicEvidenceCount":301,"organizations":1108,"bestGrade":71,"headline":1112,"lastVerified":145,"indexable":12,"euAiActTier":146,"euAiActBasis":1113},"rfp-and-proposal-response-drafting","AI for RFP, tender and sales proposal response drafting","RFP and proposal drafting","AI that helps sales and bid teams answer requests for proposal, tenders, security questionnaires and sales proposals: it breaks the request into questions and requirements, retrieves approved answers and past proposals, drafts the response and a compliance matrix, and routes open points to subject matter experts, with a proposal manager reviewing everything before submission.",[61,334,62],[319,233],[110,176,132],[1109,1110,910,1111],"GroupeActive","Industrialized Construction Group","Verdantas",{"kpi":43,"label":44,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":987,"qualifier":98,"claimant":99,"organization":1110,"vendorReported":12},"Drafting bid responses for staff to review is not listed in Annex III, and the buyer receives the seller's own document rather than interacting with an AI system, so the high risk tier and the Article 50(1) duty towards the buyer do not apply. Staff who chat with the agent must know it is an AI system, which an internal tool labelled as an AI assistant meets by design. Article 50(2) does apply to the drafting itself: the provider of a system that generates text must mark its output in a machine readable format as artificially generated, whether or not a person reviews the draft, unless the system only performs an assistive function for standard editing. A seller that uses a third party drafting tool relies on that tool's provider for the marking; a seller that builds its own agent that generates proposal text, as GroupeActive did with Witivio on Copilot Studio, can be the provider and then carries the duty itself. AI literacy under Article 4 applies in both cases, and the seller remains responsible for every statement in the submitted response.",{"slug":1115,"title":1116,"shortTitle":1117,"definition":1118,"status":19,"industries":1119,"functions":1120,"patterns":1121,"audience":65,"autonomy":90,"adoptionStage":66,"evidenceCount":34,"publicEvidenceCount":34,"organizations":1122,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1129},"inspection-prioritization","AI for risk based inspection prioritization in food safety, workplace and environmental regulation","Inspection prioritization","Models that predict which premises, operators or activities are most likely to be non compliant, so that inspectors in food safety, workplace safety, environmental and other regulation spend their visits where the risk is highest, ideally with inspectors choosing the visits and random inspections testing the model.",[251],[393,254,392],[28,27],[1123,1124,1125,1126,1127,1128],"Care Quality Commission","Driver and Vehicle Standards Agency","U.S. Environmental Protection Agency, Office of Enforcement and Compliance Assurance","Food Standards Agency","Nederlandse Arbeidsinspectie","Nederlandse Voedsel- en Warenautoriteit (NVWA)","Prioritizing inspections of businesses and premises is not a use listed in Annex III, so such a system is usually not high risk. The assessment changes when it scores natural persons, such as individual licensed professionals or sole traders, and the inspectorate acts as a law enforcement authority: assessing the risk that a person offends, or profiling persons in the detection or investigation of criminal offences, is high risk under Annex III point 6 (d) and (e), and predicting that a person will commit a criminal offence based solely on profiling is prohibited by Article 5(1)(d). GDPR applies wherever sole traders, home based businesses or named professionals are scored.",{"slug":1131,"title":1132,"shortTitle":1133,"definition":1134,"status":19,"industries":1135,"functions":1136,"patterns":1138,"audience":65,"autonomy":31,"adoptionStage":66,"evidenceCount":256,"publicEvidenceCount":256,"organizations":1139,"bestGrade":41,"headline":1146,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1148},"security-alert-triage-and-investigation","AI for security alert triage and investigation in the SOC","Security alert triage","An AI agent in the security operations centre that picks up each new alert or user reported phishing email, gathers the evidence from the SIEM, endpoint, identity and threat intelligence tools, gives a verdict with its reasoning and a draft incident summary, and closes clear false positives while an analyst approves every containment action.",[61,173,62,251,334],[1137,24],"security-operations",[26,29,89,176],[1140,1141,1142,1143,1144,1145,612],"Avanade","Federal Housing Finance Agency","Human Managed","SEP2","St. Luke's University Health Network","TÜV SÜD",{"kpi":95,"label":96,"unit":45,"n":161,"nUpTo":47,"kind":457,"value":1147,"qualifier":98,"claimant":165,"organization":165,"vendorReported":11},60,"Triage of phishing, endpoint, network and cloud alerts for an organization's own cyber defence is not listed in Annex III. Recital 55 of the AI Act says that components intended to be used solely for cybersecurity purposes should not qualify as safety components, so the agent does not fall under Annex III point 2 (critical infrastructure), and for this scope the tier is minimal. The design changes that when the agent triages identity, data loss prevention, insider risk or user behaviour alerts in a way that scores or monitors individual employees: monitoring and evaluating the behaviour of persons in a work relationship falls under Annex III point 4(b), so that scope needs its own high risk assessment before it goes live. The Article 50(1) duty to disclose AI interaction does not apply because it is obvious to a reasonably well informed analyst that they are working with an AI agent. An operator that lets AI act autonomously on network or operational technology controls should assess that design separately, and reading employees' emails and sign in data remains subject to data protection law.",{"slug":1150,"title":1151,"shortTitle":1152,"definition":1153,"status":19,"industries":1154,"functions":1155,"patterns":1156,"audience":30,"autonomy":111,"adoptionStage":66,"segment":30,"evidenceCount":301,"publicEvidenceCount":301,"organizations":1157,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1160},"settlement-fail-prediction-and-exception-management","AI for settlement fail prediction and post trade exception management","Settlement fail prediction","AI that scores each pending securities settlement instruction for its likelihood of failing, names the probable cause (unmatched instruction, wrong settlement details, lack of securities or cash), and helps operations teams work the exceptions and counterparty queries before the intended settlement date, so fewer trades fail and fewer late settlement penalties are paid.",[447,154,197],[158,393],[28,29,26,110],[731,1158,1159],"Clearstream","Euroclear","Predicting settlement fails and handling post trade exceptions between professional market participants is not a use listed in Annex III and is not a prohibited practice under Article 5, so the tier depends on how the agent communicates. While an operator reviews and sends every message, the system is minimal risk: the messages are the firm's own correspondence and the firm as deployer owes AI literacy for staff (Article 4). Once the agent sends queries or chasers to counterparty or custodian staff itself, as the playbook recommends for routine information requests, it interacts directly with natural persons and Article 50(1) requires telling the recipients they are dealing with an AI system. In both designs the provider of the text generating system must mark its output as AI generated in a machine readable format under Article 50(2). Model risk and operational resilience controls apply on top.",{"slug":1162,"title":1163,"shortTitle":1164,"definition":1165,"status":19,"industries":1166,"functions":1167,"patterns":1168,"audience":65,"autonomy":111,"adoptionStage":66,"evidenceCount":301,"publicEvidenceCount":301,"organizations":1169,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":1172},"software-vulnerability-remediation","AI for software vulnerability triage and remediation","Vulnerability remediation","AI that takes security findings from scanners, fuzzers and bug reports, filters out duplicates and false positives, reproduces and ranks the real ones, and drafts a code fix with a test for each, which a developer reviews and merges through the normal change process.",[61,62,173],[1137,24],[299,26,29],[909,1170,1171],"Labelbox","PatientPoint","Drafting and triaging code fixes for an organization's own software is not an Annex III use, and developers, not the public, interact with the system. The software being fixed remains subject to its own security and resilience rules, whoever wrote the fix.",{"slug":1174,"title":1175,"shortTitle":1176,"definition":1177,"status":19,"industries":1178,"functions":1179,"patterns":1180,"audience":30,"autonomy":90,"adoptionStage":66,"segment":127,"evidenceCount":301,"publicEvidenceCount":67,"organizations":1181,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":1184},"subrogation-opportunity-detection","AI for subrogation opportunity detection","Subrogation detection","AI that reads open and closed claims to find cases where a third party is wholly or partly liable, estimates liability and the recoverable amount under the applicable negligence and recovery rules, and sends scored recovery opportunities with their reasons to the subrogation team.",[125],[127,828],[29,28,132,89],[1182,1183],"Central Insurance","Elephant Insurance","Detecting recovery opportunities against third parties and other insurers is not listed in Annex III: point 5(c) covers only risk assessment and pricing of natural persons in life and health insurance, and the system does not decide on a natural person's access to a service. It is an internal tool that does not converse with the public or publish generated content, so the deployer transparency duties of Article 50 do not apply. Personal data in claim files, including data about the third party, is still subject to GDPR.",{"slug":1186,"title":1187,"shortTitle":1188,"definition":1189,"status":19,"industries":1190,"functions":1191,"patterns":1192,"audience":65,"autonomy":111,"adoptionStage":32,"segment":529,"evidenceCount":161,"publicEvidenceCount":161,"organizations":1193,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":1195},"supervisory-exam-response-assembly","AI for supervisory exam and information request responses","Exam response assembly","An assistant for the bank's regulatory affairs team that reads a supervisory information request or exam question, retrieves the relevant evidence, policies and prior correspondence, drafts a response for legal and compliance to approve, and tracks every commitment and remediation action through to closure.",[154,125,447,155],[392,407,254],[176,110,132,26],[1194,260],"U.S. Department of Homeland Security","Drafting regulatory correspondence for human approval is not an Annex III use. The main risks are confidentiality and accuracy, which are handled by supervisory information rules, data protection law and internal controls.",{"slug":1197,"title":1198,"shortTitle":1199,"definition":1200,"status":19,"industries":1201,"functions":1203,"patterns":1204,"audience":30,"autonomy":31,"adoptionStage":177,"segment":30,"evidenceCount":134,"publicEvidenceCount":301,"organizations":1205,"bestGrade":41,"headline":1208,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":1209},"supplier-invoice-processing","AI for supplier invoice processing in accounts payable","Supplier invoice processing","AI that captures supplier invoices from any format, extracts header and line data, matches them to purchase orders and goods receipts, proposes tax and cost centre coding, flags duplicates and suspected fraud, and routes them for approval and posting, leaving only exceptions to accounts payable staff.",[61,154,251,85,1202],"energy-and-utilities",[408,406],[132,26,27,29],[789,1206,612,1207],"Kingfisher","Veolia",{"kpi":95,"label":96,"unit":45,"n":46,"nUpTo":46,"kind":48,"value":987,"qualifier":98,"claimant":51,"organization":1206,"vendorReported":11},"Processing supplier invoices is not an Annex III use case, is not a practice prohibited by Article 5 and does not involve decisions about natural persons, so it is minimal risk and the AI literacy duty of Article 4 applies. Approvers who ask questions in chat use an internal tool they know is AI; if that is not obvious to the people using it, the provider must also inform them that they are interacting with an AI system (Article 50(1)).",{"slug":1211,"title":1212,"shortTitle":1213,"definition":1214,"status":19,"industries":1215,"functions":1216,"patterns":1217,"audience":65,"autonomy":111,"adoptionStage":32,"evidenceCount":301,"publicEvidenceCount":301,"organizations":1218,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":1221},"support-knowledge-article-generation","AI for support knowledge article generation and maintenance","Knowledge article generation","AI that drafts knowledge base articles from resolved tickets, cases and conversations, detects questions the knowledge base does not answer and articles that are outdated or contradict each other, and proposes new or revised articles for a knowledge owner to review and publish.",[61,251,554],[233,128,24],[110,89,29],[1060,413,1219,1220],"U.S. National Science Foundation","Rivian","Drafting internal or public help content that a person reviews and publishes is not a prohibited practice under Article 5 and is not listed in Annex III, so it is minimal risk. The articles are not a direct AI interaction, and the Article 50(4) disclosure for AI generated text published to inform the public does not apply where a person reviews the text and holds editorial responsibility. The Article 50 transparency duties do apply to chatbots that later answer customers from the articles.",{"slug":1223,"title":1224,"shortTitle":1225,"definition":1226,"status":19,"industries":1227,"functions":1228,"patterns":1229,"audience":30,"autonomy":90,"adoptionStage":66,"evidenceCount":161,"publicEvidenceCount":161,"organizations":1230,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1232},"tax-compliance-risk-scoring","AI for tax compliance risk scoring and audit selection","Tax compliance risk scoring","Models that score tax returns, taxpayers and transactions for the risk of error, underreporting or fraud, so that a tax administration spends its audit and compliance capacity where the risk is highest, with an officer deciding every compliance action and the selection itself monitored for fairness.",[251],[393,254,157],[28,27],[1231,437,413],"Belastingdienst","Risk selection for administrative tax audits is not listed in Annex III, and Recital 59 says systems used by tax and customs authorities in administrative proceedings should not be treated as high risk law enforcement systems. Use in criminal tax investigations (Annex III point 6, law enforcement), or evaluating the eligibility of natural persons for public assistance benefits run through the tax system (Annex III point 5(a)), can make it high risk. When individuals are scored in administrative tax work, the GDPR applies, including its profiling rules (Member States may restrict some rights for taxation matters under Article 23). Article 22 applies when a decision with legal or similarly significant effect is taken solely by the model. Criminal investigations fall outside the GDPR and under the Law Enforcement Directive (EU) 2016/680 instead.",{"slug":1234,"title":1235,"shortTitle":1236,"definition":1237,"status":19,"industries":1238,"functions":1239,"patterns":1240,"audience":65,"autonomy":111,"adoptionStage":66,"segment":529,"evidenceCount":301,"publicEvidenceCount":301,"organizations":1241,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":1243},"vendor-due-diligence","AI for third party and vendor risk due diligence","Vendor due diligence","AI that reviews a vendor's security questionnaires, SOC and assurance reports, contracts and model documentation against the organization's control requirements, researches the vendor's ownership, sanctions, financial health and adverse media, drafts the risk assessment for a human to approve and keeps the register of material service providers current with ongoing monitoring.",[61,154,125,251,155],[406,393,392],[132,176,26,89],[802,413,340,1242],"U.S. Trade and Development Agency","Assessing organizations as vendors is not an Annex III use. If assessments score individual natural persons, such as sole traders, check the design against Annex III and data protection rules. The EU AI Act also shapes what to ask AI vendors, since providers of high risk systems carry specific obligations.",{"slug":1245,"title":1246,"shortTitle":1247,"definition":1248,"status":19,"industries":1249,"functions":1250,"patterns":1251,"audience":30,"autonomy":111,"adoptionStage":177,"evidenceCount":134,"publicEvidenceCount":134,"organizations":1252,"bestGrade":41,"headline":1258,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1259},"customer-feedback-analysis","AI for voice of the customer and feedback analysis","Customer feedback analysis","AI that reads every piece of free text customer feedback, such as survey verbatims, NPS comments, reviews, social posts, chat and call transcripts, and turns it into themes, sentiment, drivers and suggested actions that a named owner can act on, so the organization hears all of its customers instead of a sample.",[61,85,251,404],[128,512,87],[29,89,235],[1253,1254,1255,1256,1257],"U.S. Department of Housing and Urban Development","Majid Al Futtaim Retail","Mattel","SBF Group","U.S. Social Security Administration",{"kpi":142,"label":143,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":116,"qualifier":98,"claimant":99,"organization":1256,"vendorReported":12},"Classifying and summarizing text feedback is minimal risk. The tier changes if the system infers emotions from customers' voices or faces in calls or video: emotion recognition based on biometric data is listed as high risk in Annex III point 1(c) and triggers the Article 50(3) duty to inform the people exposed. Analysing feedback from employees to evaluate individual workers moves it towards Annex III point 4(b), and emotion recognition in the workplace is prohibited by Article 5(1)(f), except for medical or safety reasons.",{"slug":1261,"title":1262,"shortTitle":1263,"definition":1264,"status":19,"industries":1265,"functions":1266,"patterns":1267,"audience":30,"autonomy":111,"adoptionStage":66,"segment":160,"evidenceCount":161,"publicEvidenceCount":67,"organizations":1268,"bestGrade":71,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1270},"portfolio-reporting-and-commentary","AI generated client portfolio reports and commentary","Portfolio commentary","AI that drafts each client's periodic portfolio commentary and report narrative (performance, attribution, what drove returns, positioning and outlook) in plain language and in the client's language, where every figure comes from the portfolio system of record and a reviewer approves the text before delivery.",[197,154],[87,128,158],[110,89,216],[396,1269],"Quilter","Drafting client reports for human review is not listed in Annex III and is not a practice prohibited by Article 5, so the tier turns on the firm's role under Article 50. A firm that deploys a third party generator (for example a feature of its portfolio platform) for private client reports has no Article 50 duty: the Article 50(4) disclosure duty covers AI generated text published to inform the public on matters of public interest, which private client reports are not, and it lapses anyway after human review under editorial responsibility. For that firm the tier is minimal. A firm that builds the generating system or places it on the market under its own name is a provider under Article 50(2) and must mark the synthetic text in a machine readable format; drafting whole commentaries goes beyond the exemption for an assistive function for standard editing, so for that firm the tier is limited.",{"slug":1272,"title":1273,"shortTitle":1274,"definition":1275,"status":19,"industries":1276,"functions":1277,"patterns":1278,"audience":65,"autonomy":90,"adoptionStage":66,"evidenceCount":301,"publicEvidenceCount":301,"organizations":1279,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1283},"internal-talent-marketplace-matching","AI internal talent marketplace for matching employees to projects, roles and mentors","Internal talent marketplace","An internal platform that uses AI to infer employees' skills and interests and recommend short term projects, open roles, mentors and learning to them, while showing managers which employees fit an opportunity, so that work is staffed from inside before hiring or contracting externally.",[61,404,155,251],[108],[376,28],[1280,1096,1281,1282],"Federal Bureau of Prisons","Schneider Electric","Unilever","Annex III point 4 lists AI used for the recruitment or selection of natural persons (4(a)) and AI used to make decisions affecting promotion, or to allocate tasks based on individual behaviour, personal traits or characteristics (4(b)). A marketplace that ranks employees for internal roles or allocates projects on the basis of inferred traits is therefore high risk. Recommending learning content or mentors to an employee who chooses freely is usually not. Deployers of the high risk part must inform workers' representatives and the affected employees before use (Article 26).",{"slug":1285,"title":1286,"shortTitle":1287,"definition":1288,"status":19,"industries":1289,"functions":1290,"patterns":1291,"audience":65,"autonomy":90,"adoptionStage":177,"segment":202,"evidenceCount":34,"publicEvidenceCount":34,"organizations":1292,"bestGrade":41,"headline":165,"lastVerified":52,"indexable":12,"euAiActTier":146,"euAiActBasis":1295},"wealth-advisor-knowledge-assistant","AI knowledge assistant for wealth advisors and relationship managers","Advisor knowledge assistant","A conversational assistant that answers a wealth advisor's or relationship manager's questions in seconds from the firm's own research, house view, product documentation and policies, with every answer linked to the source document so the advisor can check it before using it with a client.",[197,154],[233,319,128],[176,130],[181,453,516,396,1293,1294],"UBS","Yes Bank","Article 50(1) requires that people who interact directly with an AI system are informed of it, unless this is obvious from the context, as it usually is for an internal assistant labelled as AI; Article 50(2) requires providers of systems that generate text to mark the output as AI generated in a machine readable way. Helping advisors find information is not an Annex III use and not a prohibited practice under Article 5. It would become high risk only if the system were used to evaluate the creditworthiness of clients (point 5(b)) or to evaluate or make decisions about advisors (point 4(b)). If the assistant were opened to clients, they would have to be told they are dealing with AI.",{"slug":1297,"title":1298,"shortTitle":1299,"definition":1300,"status":19,"industries":1301,"functions":1302,"patterns":1303,"audience":65,"autonomy":111,"adoptionStage":66,"evidenceCount":301,"publicEvidenceCount":301,"organizations":1304,"bestGrade":41,"headline":1307,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1309},"legal-research-and-drafting-assistant","AI legal research and drafting assistant for lawyers","Legal research and drafting","A generative AI assistant for lawyers in firms, legal departments and public bodies that finds and summarises case law, legislation and internal know how, answers legal questions with citations and drafts first versions of memos, briefings, letters and filings, which a lawyer verifies and signs off.",[334,61,251],[407,233],[176,110,89,132],[1305,1306,802,955],"A&O Shearman","Ashurst Perkins Coie",{"kpi":95,"label":96,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":1308,"qualifier":750,"claimant":51,"organization":1306,"vendorReported":11},45,"Research and drafting support for lawyers in firms and companies is not listed in Annex III, so it is normally minimal risk with AI literacy duties. Annex III point 8(a) makes it high risk when a judicial authority, or someone on its behalf, uses AI to research and interpret facts and the law and to apply the law to a concrete set of facts, or when it is used in a similar way in alternative dispute resolution, so a deployment for courts, tribunals or arbitration needs its own classification.",{"slug":1311,"title":1312,"shortTitle":1313,"definition":1314,"status":19,"industries":1315,"functions":1316,"patterns":1317,"audience":65,"autonomy":111,"adoptionStage":177,"segment":202,"evidenceCount":34,"publicEvidenceCount":34,"organizations":1318,"bestGrade":41,"headline":1321,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":1323},"client-meeting-notes-and-crm-update","AI meeting notes and CRM update for wealth advisors","Advisor meeting notes","An AI notetaker for wealth advisors that turns a client advice meeting, recorded with the client's consent, into the file note, follow up message and CRM record the firm needs to evidence its advice; unlike a general meeting summarizer, its output becomes part of the regulated client record. It drafts a structured note with the client's goals, circumstances, decisions and action items, and writes it into the CRM once the advisor has approved it.",[197,154],[319,392,158],[89,235,26,110],[181,1319,396,1269,164,1320],"Commerzbank","UniSuper",{"kpi":95,"label":96,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":1322,"qualifier":98,"claimant":99,"organization":164,"vendorReported":12},15,"Transcribing and summarizing meetings for an employee is not a use listed in Annex III, and the advisor reviews every note before it is filed or sent. The tier would change if the tool inferred emotions: emotion recognition is high risk under Annex III point 1(c), and inferring the emotions of employees at work is prohibited under Article 5(1)(f). Both stay out of scope.",{"slug":1325,"title":1326,"shortTitle":1327,"definition":1328,"status":19,"industries":1329,"functions":1330,"patterns":1331,"audience":65,"autonomy":111,"adoptionStage":177,"evidenceCount":134,"publicEvidenceCount":134,"organizations":1332,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1336},"meeting-summarization-and-action-items","AI meeting summarization and action items","Meeting summaries and action items","AI that summarizes internal and operational meetings, such as team, project, board and case meetings: it transcribes an online or in person meeting with the participants' knowledge and produces a summary, decisions and action items with owners and dates for the organizer to check and share. It is the general purpose tool; client advice meetings and sales calls, which feed a regulated record or a sales pipeline, have their own pages.",[61,251,62,334],[233,158],[89,235],[1333,1334,1335,1097,284],"U.S. Department of Labor","Ministry of Justice","Softcat","Transcribing and summarizing meetings for the participants is minimal risk. It becomes high risk under Annex III point 4(b) if transcripts are analysed to monitor or evaluate individual workers' performance or behaviour, and inferring participants' emotions from their voices or faces at work is prohibited by Article 5(1)(f). Recording and transcription also need a lawful basis and clear information to participants under GDPR.",{"slug":1338,"title":1339,"shortTitle":1340,"definition":1341,"status":19,"industries":1342,"functions":1343,"patterns":1344,"audience":65,"autonomy":111,"adoptionStage":66,"segment":1345,"evidenceCount":67,"publicEvidenceCount":67,"organizations":1346,"bestGrade":41,"headline":1349,"lastVerified":76,"indexable":12,"euAiActTier":77,"euAiActBasis":1350},"ai-drug-discovery-platform","AI native platform for drug target discovery and molecule design","AI drug discovery platform","An AI native research platform that prioritizes disease targets from biological data, generates and optimizes candidate drug molecules computationally, and predicts their properties before a chemist synthesizes and tests them, so a pharmaceutical or biotech company reaches a validated preclinical candidate with far fewer molecules made and tested than a conventional medicinal chemistry program.",[510],[158,87],[28,110],"drug discovery",[1347,1348],"Insilico Medicine","Recursion Pharmaceuticals",{"kpi":43,"label":44,"unit":45,"n":67,"nUpTo":47,"kind":48,"value":1147,"qualifier":750,"claimant":51,"organization":1347,"vendorReported":11},"Target scoring and molecule generation are not a safety component of an Annex I product and are not one of the Annex III high risk areas (Article 6), so they do not become high risk on that route. Insofar as the platform and its training are themselves scientific research and development, activity that stays there falls outside the Regulation entirely under the Article 2(6) research exclusion. The resulting drug candidate is separately regulated as a medicine, not as an AI system, through the normal pharmaceutical approval pathway; conventional preclinical and clinical testing validates the AI's outputs before anything reaches a patient.",{"slug":1352,"title":1353,"shortTitle":1354,"definition":1355,"status":19,"industries":1356,"functions":1357,"patterns":1358,"audience":65,"autonomy":90,"adoptionStage":66,"segment":202,"evidenceCount":134,"publicEvidenceCount":134,"organizations":1359,"bestGrade":41,"headline":1360,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1361},"next-best-action-for-advisors","AI next best action prompts for wealth advisors","Advisor next best action","An AI engine for wealth advisors, not customers, that scans an advisor's whole book and surfaces a short, ranked list of client specific prompts, such as idle cash, a maturing deposit, a concentration to review, a life event or an early sign of attrition, each with the reasoning and data behind it, for the advisor to act on or dismiss.",[197,154],[319,512,87],[376,28,110],[352,453,516,396,1293],{"kpi":187,"label":188,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":987,"qualifier":98,"claimant":51,"organization":1293,"vendorReported":11},"Ranking investment and service prompts for an advisor is not listed in Annex III. It becomes high risk if the system evaluates the creditworthiness of natural persons, for example to decide which clients are offered lending (Annex III point 5(b)), so keep credit decisions out of the prompt engine. It is also high risk if the system itself is used to monitor or evaluate advisors' performance and behaviour, for example by scoring or ranking advisors on how they act on prompts (Annex III point 4(b)), so keep adoption reporting separate from performance management.",{"slug":1363,"title":1364,"shortTitle":1365,"definition":1366,"status":19,"industries":1367,"functions":1368,"patterns":1369,"audience":30,"autonomy":111,"adoptionStage":32,"segment":160,"evidenceCount":301,"publicEvidenceCount":161,"organizations":1370,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1372},"portfolio-drift-monitoring-and-rebalancing","AI portfolio drift monitoring and rebalancing proposals","Drift and rebalancing","Continuous monitoring of every client portfolio against its mandate or model, which detects drift beyond agreed bands and prepares a tax aware, low turnover rebalancing proposal with its rationale for an advisor or portfolio manager to approve before any trade is placed.",[197,154],[158,393,87],[27,26,28,110],[396,1371,353],"SimCorp","Monitoring portfolios and proposing trades for human approval is not listed in Annex III and is not a prohibited practice under Article 5, so the tier turns on the firm's role under Article 50. A firm that builds or brands the rationale writer in house is a provider under Article 50(2) and must mark the generated text in a machine readable format: drafting a rationale for the drift and the proposed trades goes beyond the exemption for an assistive function for standard editing, so for that firm the tier is limited. Article 50(1) also applies once the rationale reaches the client, as this page's own implementation step allows. A firm that only deploys a third party feature for internal approver use has no Article 50 duty, and for that firm the tier is minimal. Investment conduct rules such as MiFID II suitability and best execution still apply to the resulting trades.",{"slug":1374,"title":1375,"shortTitle":1376,"definition":1377,"status":19,"industries":1378,"functions":1380,"patterns":1381,"audience":65,"autonomy":90,"adoptionStage":66,"segment":1382,"evidenceCount":67,"publicEvidenceCount":67,"organizations":1383,"bestGrade":41,"headline":165,"lastVerified":76,"indexable":12,"euAiActTier":53,"euAiActBasis":1386},"freight-rail-rolling-stock-predictive-maintenance","AI predictive maintenance for freight rail rolling stock","Rail rolling stock predictive maintenance","Machine vision and machine learning that inspect freight railcar wheels, bearings and other running gear as trains pass wayside sensors and camera portals at track speed, learn what a healthy wheel or a healthy reading looks like, and flag the ones that need attention before a crack, an overheating bearing or a worn wheel causes a service failure or a derailment.",[1379],"logistics-and-transportation",[158,1046],[321,27,28],"mechanical-and-safety",[1384,1385],"BNSF Railway","Norfolk Southern","A system that flags a wheel or railcar for a qualified inspector to confirm is advisory and usually minimal risk. Under Article 6(1) it is high risk when both conditions hold: the same detection logic is built into a safety component of rolling stock or track equipment (or is itself such a product) covered by Directive (EU) 2016/797 on the interoperability of the rail system, which sits in Annex I Section B, for example if a flag were wired to trigger an automatic stop or speed restriction without a human check, and that directive requires a third party conformity assessment of the product. Under Article 2(2), as amended by Regulation (EU) 2026/1744, a high risk system of that kind is not subject to the full AI Act: only Article 6(1), Article 60a and Articles 102 to 112 apply directly, and Articles 57, 58 and 59 apply only so far as the high risk requirements have been integrated into the interoperability directive. The substantive high risk requirements reach the system through that directive instead, which Article 106 of the AI Act amends to require rail delegated and implementing acts to take those requirements into account.",{"slug":1388,"title":1389,"shortTitle":1390,"definition":1391,"status":19,"industries":1392,"functions":1393,"patterns":1394,"audience":65,"autonomy":90,"adoptionStage":66,"segment":1395,"evidenceCount":161,"publicEvidenceCount":161,"organizations":1396,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1399},"industrial-asset-predictive-maintenance","AI predictive maintenance for industrial and energy assets","Industrial predictive maintenance","Machine learning that learns the normal behaviour of industrial and energy equipment from sensor and process data, flags early signs of degradation weeks or months before a failure, and turns them into prioritised maintenance work, so plants and utilities plan repairs instead of reacting to breakdowns.",[1202,404],[158,1046],[27,28],"asset-management",[1397,560,1398],"Duke Energy","Shell","A system that advises engineers on the condition of equipment is usually minimal risk. Annex III point 2 lists AI systems intended as safety components in the management and operation of critical digital infrastructure, road traffic and the supply of water, gas, heating or electricity; if predictive maintenance acts on protection or control in a utility network, it can become high risk. Article 6(1) can also apply when the AI is a safety component of machinery or another product covered by Annex I legislation and that product must undergo a third party conformity assessment.",{"slug":1401,"title":1402,"shortTitle":1403,"definition":1404,"status":19,"industries":1405,"functions":1406,"patterns":1407,"audience":65,"autonomy":90,"adoptionStage":177,"segment":1408,"evidenceCount":67,"publicEvidenceCount":67,"organizations":1409,"bestGrade":41,"headline":1412,"lastVerified":76,"indexable":12,"euAiActTier":117,"euAiActBasis":1414},"radiology-worklist-triage","AI prioritization of radiology and imaging worklists","Radiology worklist triage","An AI system that analyzes a medical image immediately after a scan, flags time sensitive findings such as a brain bleed, a stroke causing large vessel occlusion or a pulmonary embolism, and reorders the radiologist's worklist and notifies the care team so the most urgent cases are read and acted on first, while a radiologist confirms every finding before it changes a patient's treatment.",[173],[158],[321,29,27],"emergency and inpatient imaging",[1410,1411],"Adventist Health + Rideout","Sheba Medical Center",{"kpi":43,"label":44,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":1413,"qualifier":750,"claimant":99,"organization":1410,"vendorReported":12},44,"Article 6(1) and Annex I: software that analyzes a medical image to detect or prioritize a disease finding is itself, or is a safety component of, a device in scope of the EU Medical Device Regulation, and typically needs a notified body conformity assessment as software as a medical device (the FDA's AI Enabled Medical Device List shows US market authorization for devices in this category, listing authorized stroke triage devices from Viz.ai and Aidoc's BriefCase triage devices), which makes it high risk under the EU AI Act regardless of Annex III. The radiologist's own diagnostic read stays a human decision; the AI narrows and reorders the queue. Annex I high risk classification under Article 6(1) applies from 2 August 2028 (Article 113(c)); until then, Article 4 (AI literacy obligations) and Article 5 (prohibited practices), which bind the hospital as a deployer, already apply.",{"slug":1416,"title":1417,"shortTitle":1418,"definition":1419,"status":19,"industries":1420,"functions":1421,"patterns":1422,"audience":30,"autonomy":31,"adoptionStage":66,"evidenceCount":134,"publicEvidenceCount":134,"organizations":1423,"bestGrade":71,"headline":1429,"lastVerified":145,"indexable":12,"euAiActTier":117,"euAiActBasis":1433},"call-quality-and-compliance-monitoring","AI quality and compliance monitoring of every customer interaction","Call quality and compliance","Automated quality assurance that transcribes and scores every customer interaction, voice and chat, against the organization's own rubric, checking required disclosures and script adherence, flagging conduct and mis selling risk, and surfacing coaching opportunities, instead of the small sample a human QA team can review.",[61,154,125,1202,21,85],[128,392,158],[235,29,89],[1424,1425,1426,1427,1428],"British Gas","Central Bank","DoorDash","Oportun","VitalityHealth",{"kpi":1430,"label":1431,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":1432,"qualifier":750,"claimant":99,"organization":1424,"vendorReported":12},"quality-score-uplift","Quality score uplift",10,"Scoring individual agents' interactions to monitor and evaluate their performance and behaviour falls under Annex III point 4(b), employment and worker management. The Article 6(3) exception does not apply where the system profiles natural persons. Inferring agents' emotions is prohibited under Article 5(1)(f), except for medical or safety reasons. Inferring customers' emotions from their voice is emotion recognition on biometric data: high risk under Annex III point 1(c), and Article 50(3) requires informing the people exposed to it. Analytics that only aggregate interaction themes without evaluating individuals can fall outside the high risk category.",{"slug":1435,"title":1436,"shortTitle":1437,"definition":1438,"status":19,"industries":1439,"functions":1440,"patterns":1441,"audience":65,"autonomy":31,"adoptionStage":66,"segment":557,"evidenceCount":161,"publicEvidenceCount":161,"organizations":1443,"bestGrade":41,"headline":1446,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1447},"production-line-quality-inspection","AI quality inspection on the production line","Production quality inspection","AI that inspects every unit on a production line, from camera images, sound or machine process data, to find defects, missing parts and wrong variants in real time, and routes the few anomalies it flags to a quality inspector instead of relying on manual sampling at the end of the line.",[404,554],[158],[321,27,1442,26],"synthetic-data-generation",[1444,559,1445],"Audi","Pegatron",{"kpi":73,"label":74,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":256,"qualifier":98,"claimant":99,"organization":1445,"vendorReported":12},"Inspecting products is not an Annex III use, so a system that only judges parts, welds or assemblies is usually minimal risk. Two designs change that. Under Article 6(1) it is high risk when both conditions hold: it is a safety component of a product (or itself a product) covered by the Union harmonisation legislation in Annex I, and that law requires a third party conformity assessment of the product. For a production line the relevant product laws are the Machinery Regulation (EU) 2023/1230 and, for cars, the vehicle type approval regulations. Since the Digital Omnibus on AI, Regulation (EU) 2026/1744, moved the Machinery Regulation into Annex I Section B, where the vehicle type approval regulations already sat. Article 6(1) still classifies such a safety component as high risk, but under Article 2(2) only Article 6(1), Article 60a and Articles 102 to 112 of the AI Act apply directly. The requirements reach the system through the sectoral law instead: delegated acts amending Annex III of the Machinery Regulation, and type approval for vehicles. The AI Act rules for Article 6(1) high risk systems apply from 2 August 2028. An inspection system on the assembly line is usually not a safety component of the product it inspects. If it monitors and evaluates the performance and behaviour of individual workers, for example by scoring who made an assembly error, it falls under Annex III point 4(b) and is high risk. Keep the output about the unit, not the person.",{"slug":1449,"title":1450,"shortTitle":1451,"definition":1452,"status":19,"industries":1453,"functions":1454,"patterns":1455,"audience":65,"autonomy":90,"adoptionStage":66,"segment":1456,"evidenceCount":301,"publicEvidenceCount":67,"organizations":1457,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":146,"euAiActBasis":1459},"regulatory-horizon-scanning","AI regulatory horizon scanning and obligation mapping","Regulatory horizon scanning","An AI system that continuously reads publications from the regulators and standard setters an organization answers to, classifies each item by relevance and urgency, breaks new rules into individual obligations and maps them to the internal policies and controls that meet them, so compliance owners see what changed and where the gaps are.",[61,154,125,155,197,510,251],[392,407,393],[29,132,176,89,26],"compliance",[1458,412],"Financial Conduct Authority","An internal tool that monitors and classifies regulatory publications for staff makes no decisions about natural persons, so it is not listed in Annex III and is not a prohibited practice under Article 5. Staff know they are using an AI tool and its summaries are not published to the public, so the Article 50 duties to inform users and to disclose published generated text add little for the deploying organization. Article 50(2) still requires the provider of a system that generates text to mark its output, in a machine readable format, as AI generated: usually the vendor, but an organization that builds its own summariser can itself be that provider, which is what puts this use case at the limited tier rather than minimal. Beyond this and AI literacy (Article 4), no specific obligations apply. General model risk and third party rules still apply.",{"slug":1461,"title":1462,"shortTitle":1463,"definition":1464,"status":19,"industries":1465,"functions":1466,"patterns":1467,"audience":65,"autonomy":90,"adoptionStage":66,"evidenceCount":161,"publicEvidenceCount":161,"organizations":1468,"bestGrade":41,"headline":1471,"lastVerified":52,"indexable":12,"euAiActTier":53,"euAiActBasis":1475},"conversation-roleplay-training","AI roleplay training for customer conversations","Conversation roleplay training","A training simulator in which generative AI plays a realistic customer, by voice or text, so service, sales and crisis staff can rehearse difficult conversations as often as they need before they handle live ones, and receive structured feedback against the organization's own standards.",[61,154,125,21,251,173],[108,128,319],[130,131,110],[181,1469,1470],"GoHealth","U.S. Department of Veterans Affairs",{"kpi":1472,"label":1473,"unit":45,"n":46,"nUpTo":47,"kind":48,"value":1474,"qualifier":98,"claimant":99,"organization":1469,"vendorReported":12},"conversion-rate-uplift","Conversion uplift",21,"Used only for practice and feedback, the simulator is limited risk. Article 50 requires that people know they are interacting with AI unless that is obvious from the context, as it usually is in a training session, and the provider must mark synthetic voice or text output as AI generated in a machine readable format. It becomes high risk under Annex III point 4(b) if its scores are used to evaluate the performance of workers or to decide on their promotion or termination, and can fall under point 3(b) when a vocational training institution uses it to evaluate learning outcomes. Inferring trainees' emotions from voice or face in the workplace is prohibited under Article 5(1)(f), except for medical or safety reasons.",{"slug":1477,"title":1478,"shortTitle":1479,"definition":1480,"status":19,"industries":1481,"functions":1482,"patterns":1483,"audience":65,"autonomy":111,"adoptionStage":66,"evidenceCount":301,"publicEvidenceCount":301,"organizations":1484,"bestGrade":71,"headline":1488,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1489},"sales-call-coaching-and-crm-update","AI sales call coaching and CRM update","Sales call coaching and CRM update","AI for sales teams that analyses sales calls and meetings against the team's own sales method to coach sellers and their managers, and writes the call summary, next steps and opportunity updates into the CRM for the seller to confirm. Its purpose is winning deals and building selling skill, not the regulated advice record or general meeting notes.",[61,21,404,125],[319],[235,89,110],[1485,1486,1487,383],"Hughes Network Systems","Lumen Technologies","Sandvik Coromant",{"kpi":563,"label":564,"unit":565,"n":46,"nUpTo":46,"kind":48,"value":161,"qualifier":98,"claimant":51,"organization":1487,"vendorReported":11},"Summaries, CRM suggestions and follow up drafts that the seller reviews are not an Annex III use and are minimal risk. Using call analysis to monitor and evaluate the performance and behaviour of individual sellers, or to allocate leads to sellers based on their behaviour or personal traits, is high risk under Annex III point 4(b). Inferring sellers' emotions from their voice is prohibited in the workplace by Article 5(1)(f). Emotion recognition applied to customers' voices is high risk under Annex III point 1(c), and Article 50(3) requires deployers to inform the people exposed to it.",{"slug":1491,"title":1492,"shortTitle":1493,"definition":1494,"status":19,"industries":1495,"functions":1497,"patterns":1498,"audience":30,"autonomy":31,"adoptionStage":177,"evidenceCount":161,"publicEvidenceCount":161,"organizations":1499,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":117,"euAiActBasis":1503},"automated-scoring-of-written-responses","AI scoring of essays and written answers in assessments","Essay and written answer scoring","AI that scores students' essays and short written answers against a rubric, trained on responses scored by human raters, with human raters rescoring a sample of responses and every response the engine is unsure about. In hybrid programmes such as Texas, a human score is the score of record whenever a human scores a response.",[1496,251],"education",[158],[29,28],[1500,1501,1502],"ETS","Massachusetts Department of Elementary and Secondary Education","Texas Education Agency","Annex III point 3(b): AI systems intended to be used to evaluate learning outcomes in educational and vocational training institutions at all levels are high risk. Scoring that determines access to an institution or the level of education a student will receive is also covered by points 3(a) and 3(c). Schools and exam bodies that use such a system have the deployer obligations of Article 26.",{"slug":1505,"title":1506,"shortTitle":1507,"definition":1508,"status":19,"industries":1509,"functions":1510,"patterns":1511,"audience":30,"autonomy":31,"adoptionStage":66,"evidenceCount":301,"publicEvidenceCount":301,"organizations":1512,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":1515},"procurement-spend-classification","AI spend classification and spend analytics for procurement","Spend classification","AI that reads purchase orders, invoices, card transactions and contracts and assigns each line of spend to a category in the organization's taxonomy, and to the right supplier, so that procurement can see what is bought, from whom and where to consolidate or negotiate.",[61,251,404,173],[406,408,87],[29,132,89],[1513,413,340,1514],"U.S. General Services Administration","Veterans Health Administration","Classifying the organization's own purchase lines into categories is not listed in Annex III and is used internally by procurement staff, so no specific obligations apply beyond AI literacy. The data can still contain personal data, for example in purchasing card and expense lines, which brings GDPR duties. Using the classified card and expense lines to monitor or evaluate individual employees would move the system towards Annex III point 4 (employment and worker management) and a high risk assessment.",{"slug":1517,"title":1518,"shortTitle":1519,"definition":1520,"status":19,"industries":1521,"functions":1522,"patterns":1523,"audience":65,"autonomy":111,"adoptionStage":66,"segment":202,"evidenceCount":301,"publicEvidenceCount":301,"organizations":1524,"bestGrade":41,"headline":1525,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1527},"investment-research-summarization","AI summaries of investment research and the house view","Research summaries","An AI assistant that condenses long research reports, overnight market moves and the house view into short, sourced briefings for advisors and analysts, answers \"what is our view on X\" on demand, and adapts approved research for different client segments and languages, with every figure traced to the original research.",[197,447,154],[87,319,233],[89,176,110,216],[453,205,396,1293],{"kpi":563,"label":564,"unit":565,"n":47,"nUpTo":46,"kind":48,"value":1526,"qualifier":224,"claimant":51,"organization":205,"vendorReported":11},120,"Summarizing research for staff is not an Annex III use and is not a practice prohibited by Article 5, so the tier turns on the firm's role under Article 50. It is minimal for a purchased internal tool with no client or public facing exposure. Article 50 transparency applies when the firm builds the generating system itself, which brings the Article 50(2) duty to mark synthetic text in a machine readable format; when the assistant is offered to clients as a chatbot, which brings the Article 50(1) duty to tell them they are interacting with AI; or when AI generated text is published to inform the public on matters of public interest, which brings the Article 50(4) disclosure duty unless the text has gone through human review or editorial control and a person holds editorial responsibility for it.",{"slug":1529,"title":1530,"shortTitle":1531,"definition":1532,"status":19,"industries":1533,"functions":1534,"patterns":1535,"audience":65,"autonomy":111,"adoptionStage":66,"segment":589,"evidenceCount":67,"publicEvidenceCount":67,"organizations":1536,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":117,"euAiActBasis":1537},"life-underwriting-medical-record-summarization","AI summarization of medical evidence for life and health underwriting","Life underwriting medical summaries","AI that reads the medical evidence behind a life or health insurance application (attending physician statements, electronic health records, lab results and disclosures), turns it into a structured, cited summary of conditions, treatments and dates, and maps it to the insurer's underwriting manual so an underwriter can decide faster and more consistently.",[125],[589],[132,89,176],[379,380],"Annex III point 5(c): AI intended for risk assessment and pricing in relation to natural persons in life and health insurance. Article 6(3) exempts some purely preparatory tasks, but never a system that profiles natural persons. Extracting an applicant's health conditions and mapping them to the underwriting manual evaluates their health, which is profiling, so treat the system as high risk. Under the timeline as amended, the obligations for Annex III high risk systems apply from 2 December 2027, and Article 27 requires deployers of point 5(c) systems to assess the impact on fundamental rights before first use.",{"slug":1539,"title":1540,"shortTitle":1541,"definition":1542,"status":19,"industries":1543,"functions":1544,"patterns":1545,"audience":65,"autonomy":90,"adoptionStage":66,"evidenceCount":161,"publicEvidenceCount":161,"organizations":1546,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":117,"euAiActBasis":1549},"emergency-call-triage-support","AI support for emergency call triage (112 and 911)","Emergency call triage support","AI that supports emergency call takers and dispatchers during 112 and 911 calls, with live transcription, translation, summaries, location cues and alerts for critical conditions such as cardiac arrest, while the call taker keeps every triage and dispatch decision.",[251,173],[253,158],[235,216,89,28],[1547,1548,983],"Baltimore City 911 (Emergency Communications)","Copenhagen Emergency Medical Services","Annex III point 5(d): AI systems intended to evaluate and classify emergency calls or to dispatch or set priority for emergency first response services (police, fire, medical aid) are high risk. Pure transcription that performs a narrow procedural or preparatory task may fall outside it under the Article 6(3) exceptions, but alerts that influence triage are in scope. An AI agent that speaks with callers directly, for example on a non emergency line, must also tell them they are interacting with AI (Article 50).",{"slug":1551,"title":1552,"shortTitle":1553,"definition":1554,"status":19,"industries":1555,"functions":1556,"patterns":1557,"audience":65,"autonomy":111,"adoptionStage":66,"evidenceCount":301,"publicEvidenceCount":301,"organizations":1558,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1560},"ai-model-inventory","AI system and model inventory with shadow AI discovery","AI model inventory","A governed register of every AI system and model an organization builds, buys or uses, with its owner, purpose, data, risk tier and approval status, kept current by AI that discovers unregistered use, reads the documentation and assembles the evidence a board, auditor or supervisor asks for.",[61,154,125,251,404],[393,392,24],[26,132,176,29],[777,1559,1282,802],"Office of Management and Budget","Minimal for a system level register of systems and owners with no monitoring of individual employees; it is not listed in Annex III and is the instrument deployers use to meet obligations such as the Article 26 duties for high risk systems and the Article 49 registration of Annex III systems in the EU database. Limited where the plain language assistant that staff and auditors query is not obviously an AI system to its users: under Article 50(1) its provider must then design it so people are told they are dealing with AI. Possibly high risk under Annex III point 4(b) on worker management if the discovery process monitors or evaluates the behavior of individual employees rather than staying at the level of systems and owners.",{"slug":1562,"title":1563,"shortTitle":1564,"definition":1565,"status":19,"industries":1566,"functions":1567,"patterns":1568,"audience":65,"autonomy":111,"adoptionStage":32,"evidenceCount":34,"publicEvidenceCount":134,"organizations":1569,"bestGrade":41,"headline":1574,"lastVerified":52,"indexable":12,"euAiActTier":77,"euAiActBasis":1576},"requirements-to-test-case-generation","AI that turns requirements into user stories, acceptance criteria and test cases","Requirements to test cases","AI that reads product requirements, specifications or recorded sessions, checks them for gaps, ambiguity and contradictions, and drafts structured user stories, acceptance criteria and test cases (for example in Gherkin) that QA engineers review, with each item traced back to the requirement it covers.",[62,251,154,61],[24],[110,132],[1570,1571,1572,1573,1470],"BrowserStack","Continental AG","LTIMindtree","National Aeronautics and Space Administration",{"kpi":43,"label":44,"unit":45,"n":46,"nUpTo":46,"kind":48,"value":1575,"qualifier":98,"claimant":99,"organization":1572,"vendorReported":12},67,"An internal assistant that drafts requirements artifacts and test cases for engineers is not listed in Annex III and does not interact with the public, so no specific obligations apply beyond AI literacy (Article 4). The system under test may itself fall under the Act.",{"slug":1578,"title":1579,"shortTitle":1580,"definition":1581,"status":19,"industries":1582,"functions":1583,"patterns":1584,"audience":133,"autonomy":111,"adoptionStage":66,"evidenceCount":178,"publicEvidenceCount":178,"organizations":1585,"bestGrade":41,"headline":165,"lastVerified":52,"indexable":12,"euAiActTier":53,"euAiActBasis":1588},"public-service-translation","AI translation and interpretation for multilingual public services","Public service translation","AI that translates government content, documents and conversations between officials and the public, in writing and in real time speech, so people can use public services in their own language, with human translators and interpreters reviewing what carries legal or safety weight.",[251],[253,128,158],[216,130,235,132],[1547,1586,1587,260,413,286,287,886],"Delaware County","European Commission","Assistants that talk with residents must tell people they are interacting with AI (Article 50(1)), and AI generated text published to inform the public on matters of public interest must be disclosed unless it has had human review under editorial responsibility (Article 50(4)). Internal translation that neither talks with people nor is published carries no specific obligation. Translation can also sit inside an Annex III process, such as examining asylum, visa or residence permit applications (point 7(c)) or evaluating emergency calls and dispatching emergency services (point 5(d)). Whether the translation component is itself high risk depends on its intended purpose (Article 6(3) exempts systems that only perform a narrow procedural task); either way it should be governed with that high risk process.",{"slug":1590,"title":1591,"shortTitle":1592,"definition":1593,"status":19,"industries":1594,"functions":1595,"patterns":1596,"audience":133,"autonomy":90,"adoptionStage":32,"evidenceCount":161,"publicEvidenceCount":161,"organizations":1597,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1601},"ai-tutor-for-students","AI tutor that coaches students through problems","AI tutor for students","An AI tutor that works with a student on course material in a conversation, asking questions and giving hints instead of handing over answers, grounded in the course content and set up by the school or teacher, with limits on use and a clear route to a human teacher.",[1496],[128],[130,176],[1598,1599,1600],"Hamilton County Schools","Harvard University","World Bank","A tutor that only converses with students falls under the transparency duty of Article 50. It becomes high risk under Annex III point 3(b) when it evaluates learning outcomes, including when those outcomes are used to steer a student's learning process, and under point 3(c) when it assesses the level of education a student should receive. Inferring students' emotions is prohibited in education institutions under Article 5(1)(f).",{"slug":1603,"title":1604,"shortTitle":1605,"definition":1606,"status":19,"industries":1607,"functions":1608,"patterns":1609,"audience":65,"autonomy":111,"adoptionStage":66,"segment":1610,"evidenceCount":67,"publicEvidenceCount":67,"organizations":1611,"bestGrade":71,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1614},"power-line-vegetation-management","AI vegetation management for power lines","Power line vegetation management","AI that analyses satellite, aerial or lidar imagery of the land along power lines to estimate where and how fast vegetation will grow into the lines or fall onto them, and turns that into a risk based trimming and hazard tree removal plan, replacing fixed trimming cycles and manual patrols.",[1202],[23,1046],[321,28],"grid",[1612,1613],"Entergy","National Grid","Annex III point 2 makes AI systems high risk when they are intended as safety components in the management and operation of the supply of electricity. Recital 55 defines such components as systems used to directly protect the physical integrity of critical infrastructure or the health and safety of persons and property. A system that only feeds a multi year trimming plan, which vegetation planners review and approve before crews act, informs maintenance rather than directly protecting the network, and is then usually minimal risk. The assessment changes when the design acts directly on protection, for example when vegetation risk scores automatically trigger fire risk protection settings or switch lines off without a person deciding; such a system should be assessed as a possible safety component. Standard GDPR duties apply where imagery shows private property or people.",{"slug":1616,"title":1617,"shortTitle":1618,"definition":1619,"status":19,"industries":1620,"functions":1621,"patterns":1622,"audience":30,"autonomy":31,"adoptionStage":66,"segment":160,"evidenceCount":46,"publicEvidenceCount":46,"organizations":1623,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":53,"euAiActBasis":1625},"dynamic-customer-risk-rating","Dynamic AML customer risk rating with machine learning","Dynamic customer risk rating","Explainable machine learning that produces the money laundering risk rating itself: it computes and continuously updates each customer's rating from due diligence data, products, geography, behaviour and screening results, and shows which factors drive the rating and when enhanced due diligence is warranted.",[154,155,197],[200,393],[28,27],[1624],"bunq","An AML customer risk rating is not listed in Annex III. Article 5(1)(d) prohibits AI risk assessments that predict whether a natural person will commit or will likely commit a criminal offence based solely on profiling of that person or on assessing their personality traits and characteristics; it exempts only AI that supports the human assessment of a person's involvement in a criminal activity, which is already based on objective and verifiable facts directly linked to a criminal activity. An AML customer risk rating built from due diligence attributes, transaction behaviour and screening results is itself an automated evaluation of a person's situation and behaviour, which is profiling under GDPR Article 4(4), and due diligence facts such as occupation, geography and products are not facts directly linked to a criminal activity, so the rating does not sit squarely inside the exemption. What keeps it a defensible AML due diligence tool rather than an offence prediction is that it does not itself accuse a person of an offence: it sets a level of scrutiny, a human analyst reviews material moves, and regulatory minimum rules sit above the model as hard constraints. A rating driven mainly by nationality or other personal attributes weakens that position further, which is why the proxy discrimination guardrail matters. If the same score is used to evaluate the creditworthiness of natural persons or to establish their credit score, that use falls under Annex III point 5(b) and is high risk, so keep the AML rating and credit decisions separate.",{"slug":1627,"title":1628,"shortTitle":1629,"definition":1630,"status":19,"industries":1631,"functions":1632,"patterns":1633,"audience":65,"autonomy":111,"adoptionStage":66,"evidenceCount":161,"publicEvidenceCount":161,"organizations":1634,"bestGrade":71,"headline":1638,"lastVerified":145,"indexable":12,"euAiActTier":77,"euAiActBasis":1639},"internal-audit-copilot","Generative AI copilot for internal audit","Internal audit copilot","A copilot for internal auditors that drafts planning memos and document request lists from prior audits, summarises large evidence sets, builds risk and control matrices from policies and process documents, and drafts findings and reports, with every statement traceable to its evidence and a qualified auditor accountable for every conclusion.",[61,154,125,251,447,197],[393,392,408],[176,89,110,132,27],[1635,1636,1637],"Banco Bradesco","British Columbia Investment Management Corporation","XP Inc.",{"kpi":241,"label":242,"unit":45,"n":67,"nUpTo":47,"kind":48,"value":749,"qualifier":98,"claimant":99,"organization":1635,"vendorReported":12},"An internal drafting and analysis assistant for auditors that makes no decisions about natural persons. It would need reassessment if used to evaluate individual employees' behaviour or performance, which falls under Annex III point 4(b).",{"slug":1641,"title":1642,"shortTitle":1643,"definition":1644,"status":19,"industries":1645,"functions":1646,"patterns":1647,"audience":133,"autonomy":1074,"adoptionStage":66,"segment":1648,"evidenceCount":161,"publicEvidenceCount":161,"organizations":1649,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":146,"euAiActBasis":1652},"in-car-ai-voice-assistant","Generative AI voice assistant in the car","In car voice assistant","A voice assistant built into the vehicle that uses a large language model to hold a natural conversation with the driver and passengers, controls comfort, navigation and media functions, answers questions about the car and the world, and keeps the vehicle's own commands and data under the car maker's control.",[554],[128,349],[131,130,176,26],"connected-car",[559,1650,1651],"Mercedes-Benz Group","Volkswagen","Article 50(1): people must be informed that they are interacting with an AI system unless that is obvious from the context. Article 50(2): synthetic audio output must be marked as artificially generated. A cabin assistant for comfort, media, navigation and knowledge questions is not an Annex III use. It would move towards the high risk regime if it became a safety component of the vehicle: vehicle type approval legislation is listed in Annex I Section B, and under Article 2(2) the high risk requirements reach those products only through the amendments the AI Act makes to that legislation. Keep driving and safety functions out of its reach.",{"slug":1654,"title":1655,"shortTitle":1656,"definition":1657,"status":19,"industries":1658,"functions":1659,"patterns":1660,"audience":65,"autonomy":90,"adoptionStage":66,"evidenceCount":161,"publicEvidenceCount":161,"organizations":1661,"bestGrade":41,"headline":165,"lastVerified":145,"indexable":12,"euAiActTier":146,"euAiActBasis":1664},"governed-text-to-sql-analytics","Governed text to SQL analytics assistant","Governed SQL analytics","An assistant that turns a business user's plain language question into a query against governed data, runs it under that user's own data permissions and returns the table or chart together with the SQL and the tables used, so routine ad hoc questions no longer queue for the data team.",[61,154,125,85,62,510],[87,24],[130,299,176],[1012,1662,1663],"LinkedIn","Uber Technologies","Article 50(1) requires providers to design AI systems that interact directly with people so that those people are informed they are dealing with AI, unless this is obvious from the context, as it usually is for an internal assistant. An analytics assistant that makes no decisions about people is not a prohibited practice under Article 5 and is not listed in Annex III. It would be high risk only if it were intended for an Annex III purpose, such as assessing the creditworthiness of natural persons (point 5(b)).",1790598319082]