[{"data":1,"prerenderedAt":628},["ShallowReactive",2],{"uc-production-line-quality-inspection":3,"uc-regulations":418},{"useCase":4,"evidence":208,"blitsAiDeployments":299,"benchmarks":300,"indicative":317,"related":320,"indexability":416,"includeUnpublished":214},{"title":5,"shortTitle":6,"seoTitle":7,"metaDescription":8,"status":9,"definition":10,"aliases":11,"industries":17,"functions":20,"patterns":22,"channels":27,"audience":30,"autonomy":31,"adoptionStage":32,"segment":33,"problem":34,"problemStats":35,"howItWorks":36,"valueDrivers":37,"kpis":42,"indicativeValue":49,"macroEstimates":84,"feasibility":85,"implementation":99,"risk":145,"blitsAi":190,"faq":192,"related":202,"datePublished":203,"dateModified":203,"lastVerified":203,"changelog":204,"slug":207},"AI quality inspection on the production line","Production quality inspection","AI quality inspection for manufacturing lines","AI checks each unit by camera, sound or sensor. In 2023 Audi reported about 1.5 million welds analysed per shift; NVIDIA cites a 67% lower defect rate at Pegatron.","published","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.",[12,13,14,15,16],"AI visual inspection","AI defect detection","machine vision quality control","automated optical inspection with deep learning","acoustic quality inspection",[18,19],"manufacturing","automotive",[21],"operations",[23,24,25,26],"computer-vision","anomaly-detection","synthetic-data-generation","agentic-workflow",[28,29],"internal-tools","api","employee-facing","supervised-agent","early-adopters","production","Classic quality control samples. An inspector checks a share of the units, technicians test a\nrandom selection of welds with ultrasound, a specialist walks round the finished product at the\nend of the line. Defects that fall between samples travel on to the next station, to the customer\nor into a warranty claim, further away from the station that caused them.\n\nRule based machine vision suits simple, stable parts, but every new variant, lighting change or\ndefect type means writing and tuning new rules. The shift is to learning models that check every unit, from images, sound or the process data a\nmachine already produces, and send people only the cases that look wrong.",[],"1. **Capture every unit.** Cameras, microphones or the machine controller record each part or\n   vehicle at the station, for example images of an assembly step, driving noise from the seats or\n   the process readings a machine logs for each joint.\n2. **Score it against what good looks like.** A model trained on labelled examples, or on normal\n   production when defects are rare, classifies the unit or scores how far it deviates from normal.\n   Synthetic defect images can fill the gap when real defects are too rare to train on.\n3. **Check it against the order.** The expected variant, parts list and assembly steps come from\n   the production system, so the model knows what this specific unit should look like.\n4. **Alert the right person at the right station.** An anomaly goes straight to the worker or\n   inspector on a smart device, with the image or clip, while the unit can still be fixed in line.\n5. **Close the loop.** Confirmed and rejected findings are logged, used to retrain the model and\n   analysed for root causes such as a drifting machine setting, so the process improves rather\n   than just the inspection.",[38,39,40,41],"risk-reduction","cost-to-serve","employee-productivity","speed",[43,44,45,46,47,48],"error-reduction","detection-rate-improvement","false-positive-reduction","cost-reduction","accuracy","interactions-handled",{"referenceOrg":50,"inputs":51,"formula":79,"currency":80,"period":81,"resultLabel":82,"caveat":83},"An assembly plant building 200,000 units a year",[52,58,65,72],{"key":53,"label":54,"low":55,"high":55,"unit":56,"note":57},"units","Units produced per year",200000,"units per year","The reference plant.",{"key":59,"label":60,"low":61,"high":62,"unit":63,"note":64},"defectRate","Share of units with a defect found late (final inspection, customer or warranty)",0.02,0.04,"fraction of units","Editorial assumption, replace with your own late defect and warranty rate.",{"key":66,"label":67,"low":68,"high":69,"unit":70,"note":71},"costPerDefect","Cost of a defect found late",150,400,"USD per defect","Editorial assumption covering rework, scrap and warranty handling. Replace with your own cost of poor quality.",{"key":73,"label":74,"low":75,"high":76,"unit":77,"note":78},"reduction","Share of late defects avoided by inspecting every unit in line",0.2,0.5,"fraction of late defects","Conservative against the benchmark on this page (NVIDIA reports a 67% decrease in defect rates on Pegatron assembly lines using its visual AI agent), because that figure is one vendor reported case in electronics assembly. The benchmark is not like for like: it measures fewer defects produced, while this input measures fewer defects escaping to late stages.","units * defectRate * costPerDefect * reduction","USD","per year","Cost of late defects avoided","Counts avoided rework, scrap and warranty cost only. It leaves out cameras, sensors, edge computing and integration, the labelling effort, any reduction in inspection staff hours and the value of fewer recalls and a better brand reputation.",[],{"complexity":86,"complexityNote":87,"dataPrerequisites":88,"integrations":93},"high","The model is rarely the hard part. The work is in stable image or signal capture on a moving line, enough labelled defects, a link to the order so the model knows the expected variant, and a validation approach that quality auditors and certification bodies accept.",[89,90,91,92],"Images, sound or process signals captured consistently per unit and station","Labelled examples of good units and of each defect class, or long runs of normal production","The expected variant, parts list and work steps per unit from the production system","Confirmed outcomes of flagged cases to retrain and to measure false alarms",[94,95,96,97,98],"Manufacturing execution system for orders, variants and unit identity","Cameras, microphones or machine controllers at the stations","Edge computing near the line for low latency scoring","Quality management system for defect records and corrective actions","Worker devices or line displays for alerts",{"steps":100,"guardrails":119,"humanInTheLoop":125,"kpisToInstrument":126,"failureModes":132},[101,104,107,110,113,116],{"title":102,"detail":103},"Pick one station with a costly, visible defect","Choose a defect that escapes today, is expensive later and can be seen, heard or measured at one station, for example a missing fastener or a weak weld. Record its current escape rate.",{"title":105,"detail":106},"Fix the capture before the model","Stabilise lighting, camera angle, microphone placement or signal logging first. A better model cannot make up for inconsistent capture.",{"title":108,"detail":109},"Build the defect library","Collect and label real defects with the quality team. When real defects are rare, train on normal production for anomaly detection or add synthetic defect images, and keep a real holdout set to test on.",{"title":111,"detail":112},"Run in shadow mode next to the current inspection","Let the model score every unit while the existing sampling continues, and compare both on catches, misses and false alarms before anyone relies on it.",{"title":114,"detail":115},"Agree the validation with quality and auditors","Document how a result is produced, the acceptance thresholds and the retraining rules, so the process survives audits and certification. Audi worked with DGQ and Fraunhofer for this.",{"title":117,"detail":118},"Go live with a human on every alert, then widen","Route anomalies to the station, measure confirmation rates per defect class, and only then reduce manual sampling or add stations, plants and suppliers.",[120,121,122,123,124],"A documented acceptance threshold per defect class, with a human decision on every flagged unit","A fallback to the previous inspection method when the model or capture is unavailable","Retraining only through change control, tested on a fixed holdout set before release","Drift monitoring on input images or signals after changes to materials, lighting or machines","Recording limited to the product and process, with worker privacy rules for any video of people","Quality inspectors confirm or reject every flagged unit and own the decision to release, rework or scrap. Quality engineers approve each new defect class, threshold and model version, and a sample of passed units is still inspected manually to measure what the model misses.",[127,128,129,130,131],"Escape rate of each defect class to later stations, the customer and warranty, before and after","False alarm rate per defect class and station","Share of flagged units confirmed as real defects","Inspection coverage (share of units inspected) and time to detection","Cost of poor quality per unit produced",[133,136,139,142],{"title":134,"detail":135},"False alarms that train people to ignore alerts","A model that flags too much gets overridden by habit. Tune thresholds per defect class and report confirmation rates to the line.",{"title":137,"detail":138},"Silent drift after a process change","A new supplier, paint batch or light fitting changes the input and accuracy drops without anyone noticing. Monitor input drift and recheck on a holdout set after every change.",{"title":140,"detail":141},"A model trained on too few real defects","Rare defects are exactly the ones that matter. Use anomaly detection or synthetic data, but always test on real defects.",{"title":143,"detail":144},"Inspection without root cause","The model catches more defects but nobody fixes the cause. Feed findings into corrective actions and machine settings.",{"euAiAct":146,"regulations":149,"guidance":154,"controls":183,"incidents":189},{"tier":147,"basis":148},"context-dependent","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.",[150,151,152,153],"eu-ai-act","gdpr","iso-42001","nist-ai-rmf",[155,161,165,169,173,179],{"title":156,"issuer":157,"region":158,"url":159,"note":160},"Article 2, scope","European Union","europe","https://artificialintelligenceact.eu/article/2/","Article 2(2), as amended by Regulation (EU) 2026/1744, says that for high risk AI systems related to products under the Annex I Section B laws, such as machinery and vehicle type approval, only Article 6(1), Article 60a and Articles 102 to 112 apply.",{"title":162,"issuer":157,"region":158,"url":163,"note":164},"Annex I, list of Union harmonisation legislation","https://artificialintelligenceact.eu/annex/1/","The product laws behind Article 6(1). Since Regulation (EU) 2026/1744, Section B lists the Machinery Regulation (EU) 2023/1230 as point 21, next to motor vehicle type approval (Regulations 2018/858 and 2019/2144), and the Machinery Directive 2006/42/EC is deleted from Section A.",{"title":166,"issuer":157,"region":158,"url":167,"note":168},"Annex III, high risk AI systems referred to in Article 6(2)","https://artificialintelligenceact.eu/annex/3/","Point 4(b) lists AI used to monitor and evaluate the performance and behaviour of workers, which matters when inspection video also shows people.",{"title":170,"issuer":157,"region":158,"url":171,"note":172},"Regulation (EU) 2023/1230 on machinery","https://eur-lex.europa.eu/eli/reg/2023/1230/oj","The EU Machinery Regulation replaces Directive 2006/42/EC and applies from 14 January 2027. Since Regulation (EU) 2026/1744 it is listed in Annex I Section B of the AI Act. It is the product law that can make an AI safety component of a machine high risk under Article 6(1), and the AI requirements for such machines are to be added to its Annex III by delegated acts that apply by 2 August 2028.",{"title":174,"issuer":175,"region":176,"url":177,"note":178},"IATF 16949 automotive quality management","International Automotive Task Force","global","https://www.iatfglobaloversight.org/","The body behind the IATF 16949 automotive quality management standard. At certified plants, plan for an AI based inspection process to be reviewed as part of the quality management system in certification audits.",{"title":180,"issuer":157,"region":158,"url":181,"note":182},"Digital Omnibus on AI, Regulation (EU) 2026/1744","https://artificialintelligenceact.eu/ai-act-explorer/digital-omnibus/","Entered into force on 27 July 2026. It moves machinery from Annex I Section A to Section B, limits the AI Act's direct application to machinery to the provisions in Article 2(2), and sets 2 August 2028 as the application date for Article 6(1) high risk AI systems.",[184,185,186,187,188],"Validation record per model version with holdout results per defect class","Change control for thresholds, training data and model releases","Traceability from every flagged or passed unit to the model version and input data","Data protection impact assessment where cameras capture workers","Periodic manual audit of passed units to estimate the miss rate",[],{"howToBuild":191},"The inspection model itself runs on the line, on the camera or edge platform of the manufacturer's\nchoice. Blits.ai adds the layer around it that people work with. An **AI agent** connected to a\n**SQL knowledge base** of inspection results lets quality engineers ask in plain language which\nstations, variants or shifts show rising defect rates, and a **knowledge base** with hybrid\nretrieval over work instructions and defect catalogues answers how to rework a finding.\n\n**Agentic workflows** watch the results and, when a defect pattern crosses a threshold, draft a\nquality case or corrective action and wait for **human in the loop approval** before anything is\nraised in the quality system through **custom functions**. Line staff reach the agent in\n**Microsoft Teams** or through the REST API channel on their devices, **monitors** check the\nagent's answers on a schedule, and the platform is model agnostic, with EU and UAE data residency.",[193,196,199],{"question":194,"answer":195},"How much can AI inspection reduce defects?","Published results vary widely by defect and line. NVIDIA reports that Pegatron saw a 67% decrease in defect rates on assembly lines using its visual AI agent, a vendor figure for electronics assembly. Plan for a smaller effect at first and measure escapes per defect class against your current sampling.",{"question":197,"answer":198},"Does AI inspection need cameras?","Not always. At Plant Dingolfing, BMW's Acoustic Analytics listens to driving noises through microphones on the seats as a final check before handover. Choose the signal that shows the defect most reliably at the lowest cost.",{"question":200,"answer":201},"Is AI quality inspection high risk under the EU AI Act?","Usually not when it only judges the product. It is classified as high risk if it is a safety component of a product under Annex I legislation, such as machinery or vehicles, and that law requires a third party conformity assessment. Since the Digital Omnibus on AI, machinery is in Annex I Section B, like vehicle type approval already was, so the requirements for those systems come through the Machinery Regulation (delegated acts that apply by 2 August 2028) and vehicle type approval rather than the AI Act directly. It is also high risk if it scores the performance of individual workers, which Annex III point 4(b) covers.",[],"2026-09-27",[205],{"date":203,"note":206},"First published","production-line-quality-inspection",[209,242,265],{"title":210,"useCases":211,"organization":212,"vendors":216,"summary":219,"stage":33,"year":220,"channels":221,"languages":222,"metrics":223,"outcomeDisclosed":232,"sources":233,"verification":237,"grade":239,"id":240,"organizationSlug":241},"Audi: AI quality control of resistance spot welds in car body construction",[207],{"name":213,"anonymized":214,"country":215,"region":158,"industry":19},"Audi",false,"DE",[217],{"name":213,"role":218},"in-house","After a successful pilot at its Neckarsulm site, Audi began rolling out an AI system for quality control of resistance spot welds in car body construction (project WPS Analytics). The AI analyses around 1.5 million spot welds on 300 vehicles each shift, where production staff used to check around 5,000 spot welds per vehicle with ultrasound based on random analyses, so employees can now focus on possible anomalies. Audi developed the process with the German Association for Quality (DGQ) and two Fraunhofer institutes so that it would hold up in audits and certification. It began installing the infrastructure at Audi Brussels, with Ingolstadt and the Volkswagen plant in Emden scheduled to follow, and is retraining the model for differences in weld settings.",2023,[28],[],[224],{"kpi":48,"value":225,"unit":226,"qualifier":227,"period":228,"claimant":229,"quote":230,"sourceUrl":231},1500000,"count","approximately","spot welds analysed per shift at Neckarsulm, on 300 vehicles","organization","Using artificial intelligence, Audi analyzes around 1.5 million spot welds on 300 vehicles each shift at its Neckarsulm site.","https://www.audi.com/en/press-releases/audi-begins-roll-out-of-artificial-intelligence-for-quality-control-of-spot-welds-15443",true,[234],{"url":231,"title":235,"publisher":213,"date":236},"Audi begins roll-out of artificial intelligence for quality control of spot welds","2023-06-30",{"level":238,"checkedAt":203},"source-verified","B","audi-spot-weld-quality-analytics",null,{"title":243,"useCases":244,"organization":245,"vendors":247,"summary":249,"stage":33,"year":220,"channels":250,"languages":251,"metrics":252,"outcomeDisclosed":214,"sources":253,"verification":262,"grade":239,"id":263,"organizationSlug":264},"BMW Group: AIQX camera, sensor and acoustic quality inspection in vehicle assembly",[207],{"name":246,"anonymized":214,"country":215,"region":158,"industry":19},"BMW Group",[248],{"name":246,"role":218},"BMW built AIQX (Artificial Intelligence Quality Next), an in house platform that places cameras and sensors along the assembly line, analyses their data in real time and sends line workers immediate feedback on smart devices. It checks variants and completeness and flags anomalies, and at Plant Dingolfing a sub area called Acoustic Analytics listens to driving noises through microphones on the seats as the final check before handover. BMW describes these AI quality tools as part of the production master plan for its plants worldwide. Separately, in 2025 Plant Regensburg piloted GenAI4Q, developed with Datagon AI, which builds an individual final inspection catalogue for each of the roughly 1,400 vehicles built there each day.",[28],[],[],[254,257],{"url":255,"title":256,"publisher":246},"https://www.bmwgroup.com/en/news/general/2023/aiqx.html","How AI is revolutionising production.",{"url":258,"title":259,"publisher":260,"date":261},"https://www.press.bmwgroup.com/global/article/detail/T0449729EN/artificial-intelligence-as-a-quality-booster?language=en","Artificial intelligence as a quality booster","BMW Group PressClub","2025-04-28",{"level":238,"checkedAt":203},"bmw-group-aiqx-quality-inspection","bmw-group",{"title":266,"useCases":267,"organization":268,"vendors":272,"summary":277,"stage":33,"year":278,"channels":279,"languages":280,"metrics":281,"outcomeDisclosed":232,"sources":293,"verification":296,"grade":297,"id":298,"organizationSlug":241},"Pegatron: visual AI agent that checks manual assembly steps and synthetic defect images for inspection models",[207],{"name":269,"anonymized":214,"country":270,"region":271,"industry":18},"Pegatron","TW","asia-pacific",[273,276],{"name":274,"role":275},"NVIDIA","platform",{"name":269,"role":218},"Pegatron, an electronics manufacturer with 24 sites, built an Assembly Guiding Agent on NVIDIA's video search and summarization blueprint that watches manual assembly in real time, spots missed steps such as a forgotten screw and alerts the worker, who can replay the clip and ask the agent questions. It also generates synthetic defect images from CAD drawings and historical data to train its visual inspection models, because real defect images are scarce on high yield lines. NVIDIA reports lower defect rates and labour cost per line for the assembly agent. The page is undated; the assembly agent dates from 2025, while the synthetic defect image work may be later.",2025,[28],[],[282,290],{"kpi":43,"value":283,"unit":284,"qualifier":285,"period":286,"claimant":287,"quote":288,"sourceUrl":289},67,"percent","exact","assembly lines using the Assembly Guiding Agent","vendor","By augmenting the assembly process with this AI agent, Pegatron is seeing a 7% reduction in labor costs per assembly line and a 67% decrease in defect rates.","https://www.nvidia.com/en-us/case-studies/pegatron-scales-factory-operations-with-visual-ai-digital-twins/",{"kpi":46,"value":291,"unit":284,"qualifier":285,"period":292,"claimant":287,"quote":288,"sourceUrl":289},7,"labour cost per assembly line",[294],{"url":289,"title":295,"publisher":274},"Pegatron Scales Factory Operations With Visual AI Agents, and Digital Twins",{"level":238,"checkedAt":203},"C","pegatron-visual-ai-assembly-inspection",0,[301,307,312],{"kpi":46,"label":302,"unit":284,"aggregate":232,"higherIsBetter":232,"n":303,"nUpTo":299,"median":291,"min":291,"max":291,"byClaimant":304,"vendorOnly":232,"points":305},"Cost reduction",1,{"organization":299,"vendor":303,"regulator":299,"independent":299},[306],{"evidenceId":298,"organization":269,"value":291,"qualifier":285,"claimant":287,"grade":297,"pooled":232},{"kpi":43,"label":308,"unit":284,"aggregate":232,"higherIsBetter":232,"n":303,"nUpTo":299,"median":283,"min":283,"max":283,"byClaimant":309,"vendorOnly":232,"points":310},"Error reduction",{"organization":299,"vendor":303,"regulator":299,"independent":299},[311],{"evidenceId":298,"organization":269,"value":283,"qualifier":285,"claimant":287,"grade":297,"pooled":232},{"kpi":48,"label":313,"unit":226,"aggregate":214,"higherIsBetter":232,"n":303,"nUpTo":299,"median":225,"min":225,"max":225,"byClaimant":314,"vendorOnly":214,"points":315},"Interactions handled",{"organization":303,"vendor":299,"regulator":299,"independent":299},[316],{"evidenceId":240,"organization":213,"value":225,"qualifier":227,"claimant":229,"grade":239,"pooled":232},{"low":318,"high":319},120000,1600000,[321,349,375,397],{"slug":322,"title":323,"shortTitle":324,"definition":325,"status":9,"industries":326,"functions":329,"patterns":332,"audience":335,"autonomy":31,"adoptionStage":336,"evidenceCount":291,"publicEvidenceCount":337,"organizations":338,"bestGrade":239,"headline":344,"lastVerified":203,"indexable":232},"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.",[327,328,19,18],"cross-industry","government",[21,330,331],"case-management","finance-and-accounting",[333,23,334],"document-processing","classification-and-routing","back-office","mainstream",5,[339,340,341,342,343],"Ancine","Pupuk Indonesia","U.S. Immigration and Customs Enforcement","U.S. Citizenship and Immigration Services","Volvo Group",{"kpi":47,"label":345,"unit":284,"n":303,"nUpTo":299,"kind":346,"value":347,"qualifier":348,"claimant":287,"organization":339,"vendorReported":232},"Accuracy","reported",90,"at-least",{"slug":350,"title":351,"shortTitle":352,"definition":353,"status":9,"industries":354,"functions":356,"patterns":358,"audience":30,"autonomy":361,"adoptionStage":32,"segment":362,"evidenceCount":363,"publicEvidenceCount":363,"organizations":364,"bestGrade":239,"headline":371,"lastVerified":203,"indexable":232},"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.",[355],"telecommunications",[357,21],"network-operations",[24,334,359,360,26],"rag-knowledge-assistant","summarization","copilot","network",6,[365,366,367,368,369,370],"Bell Canada","Deutsche Telekom","KDDI","Orange","Telstra","Vodafone",{"kpi":372,"label":373,"unit":284,"n":303,"nUpTo":299,"kind":346,"value":374,"qualifier":348,"claimant":229,"organization":366,"vendorReported":214},"processing-time-reduction","Cycle time reduction",95,{"slug":376,"title":377,"shortTitle":378,"definition":379,"status":9,"industries":380,"functions":381,"patterns":383,"audience":30,"autonomy":386,"adoptionStage":32,"segment":33,"evidenceCount":387,"publicEvidenceCount":387,"organizations":388,"bestGrade":239,"headline":391,"lastVerified":203,"indexable":232},"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.",[18,19],[21,382],"knowledge-management",[359,384,360,385],"conversational-agent","translation","assist",3,[246,389,390],"Georgia-Pacific","Textron Aviation",{"kpi":392,"label":393,"unit":394,"n":299,"nUpTo":303,"kind":346,"value":395,"qualifier":396,"claimant":287,"organization":390,"vendorReported":232},"time-saved-per-task","Time saved per task","minutes",18,"up-to",{"slug":398,"title":399,"shortTitle":400,"definition":401,"status":9,"industries":402,"functions":406,"patterns":409,"audience":335,"autonomy":31,"adoptionStage":410,"segment":411,"evidenceCount":387,"publicEvidenceCount":387,"organizations":412,"bestGrade":239,"headline":241,"lastVerified":203,"indexable":232},"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.",[327,403,404,405,328],"banking","insurance","capital-markets",[407,408,21],"risk-management","regulatory-compliance",[26,333,24,334],"emerging","second-line",[413,414,415],"Federal Deposit Insurance Corporation","U.S. Department of the Interior","Pension Benefit Guaranty Corporation",{"indexable":232,"reasons":417},[],[419,424,429,435,442,448,455,462,469,476,483,489,496,503,509,514,521,527,533,539,545,551,557,562,567,574,581,586,591,598,605,611,617,622],{"id":150,"label":420,"issuer":157,"region":158,"url":421,"description":422,"useCases":423,"indexable":232},"EU AI Act","https://eur-lex.europa.eu/eli/reg/2024/1689/oj","Regulation (EU) 2024/1689: risk based rules for AI systems, with obligations for high risk systems listed in Annex III and transparency duties under Article 50.",197,{"id":151,"label":425,"issuer":157,"region":158,"url":426,"description":427,"useCases":428,"indexable":232},"GDPR","https://eur-lex.europa.eu/eli/reg/2016/679/oj","General Data Protection Regulation, including Article 22 on decisions based solely on automated processing.",180,{"id":152,"label":430,"issuer":431,"region":176,"url":432,"description":433,"useCases":434,"indexable":232},"ISO/IEC 42001","ISO and IEC","https://www.iso.org/standard/81230.html","The international management system standard for AI.",110,{"id":153,"label":436,"issuer":437,"region":438,"url":439,"description":440,"useCases":441,"indexable":232},"NIST AI Risk Management Framework","NIST","north-america","https://www.nist.gov/itl/ai-risk-management-framework","Voluntary US framework to map, measure, manage and govern AI risk, with a generative AI profile.",83,{"id":443,"label":444,"issuer":157,"region":158,"url":445,"description":446,"useCases":447,"indexable":232},"dora","DORA","https://eur-lex.europa.eu/eli/reg/2022/2554/oj","Digital Operational Resilience Act for financial entities: ICT risk, incident reporting and third party risk, including AI providers.",66,{"id":449,"label":450,"issuer":451,"region":158,"url":452,"description":453,"useCases":454,"indexable":232},"uk-gdpr","UK GDPR","Information Commissioner's Office","https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/","The UK's version of the GDPR, including rules on solely automated decisions.",64,{"id":456,"label":457,"issuer":458,"region":158,"url":459,"description":460,"useCases":461,"indexable":232},"uk-consumer-duty","FCA Consumer Duty","Financial Conduct Authority","https://www.fca.org.uk/firms/consumer-duty","UK rules that require firms to deliver good outcomes for retail customers, including through automated channels.",47,{"id":463,"label":464,"issuer":465,"region":271,"url":466,"description":467,"useCases":468,"indexable":232},"mas-ai-risk-management","MAS AI risk management guidelines","Monetary Authority of Singapore","https://www.mas.gov.sg/news/media-releases/2025/mas-guidelines-for-artificial-intelligence-risk-management","Singapore's supervisory expectations for AI risk management at financial institutions, building on the FEAT principles.",36,{"id":470,"label":471,"issuer":472,"region":271,"url":473,"description":474,"useCases":475,"indexable":232},"apra-cps-230","APRA CPS 230","Australian Prudential Regulation Authority","https://www.apra.gov.au/operational-risk-management","Australian operational risk standard covering critical operations and material service providers.",25,{"id":477,"label":478,"issuer":479,"region":176,"url":480,"description":481,"useCases":482,"indexable":232},"pci-dss","PCI DSS","PCI Security Standards Council","https://www.pcisecuritystandards.org/","Security standard for any system that stores, processes or transmits cardholder data.",20,{"id":484,"label":485,"issuer":486,"region":438,"url":487,"description":488,"useCases":482,"indexable":232},"us-sr-11-7","SR 11-7 model risk management","Federal Reserve and OCC","https://www.federalreserve.gov/boarddocs/srletters/2011/sr1107.htm","US supervisory guidance on model risk management, applied by banks to AI and machine learning models.",{"id":490,"label":491,"issuer":492,"region":158,"url":493,"description":494,"useCases":495,"indexable":232},"uk-atrs","UK Algorithmic Transparency Recording Standard","UK Government","https://www.gov.uk/government/collections/algorithmic-transparency-recording-standard-hub","Mandatory transparency records for algorithmic tools used by UK central government.",16,{"id":497,"label":498,"issuer":499,"region":176,"url":500,"description":501,"useCases":502,"indexable":232},"fatf-recommendations","FATF Recommendations","Financial Action Task Force","https://www.fatf-gafi.org/en/publications/Fatfrecommendations/Fatf-recommendations.html","Global standards for anti money laundering and counter terrorist financing that national rules implement.",15,{"id":504,"label":505,"issuer":157,"region":158,"url":506,"description":507,"useCases":508,"indexable":232},"eu-amlr","EU Anti Money Laundering Regulation","https://eur-lex.europa.eu/eli/reg/2024/1624/oj","Regulation (EU) 2024/1624: the single EU rulebook for customer due diligence, beneficial ownership and suspicious transaction reporting.",14,{"id":510,"label":511,"issuer":157,"region":158,"url":512,"description":513,"useCases":508,"indexable":232},"nis2","NIS2 Directive","https://eur-lex.europa.eu/eli/dir/2022/2555/oj","Directive (EU) 2022/2555 on cybersecurity for essential and important entities, including telecom networks, energy and public administration.",{"id":515,"label":516,"issuer":517,"region":438,"url":518,"description":519,"useCases":520,"indexable":232},"us-bsa","Bank Secrecy Act","FinCEN","https://www.fincen.gov/resources/statutes-and-regulations/bank-secrecy-act","US anti money laundering law: customer due diligence, suspicious activity reports and record keeping.",13,{"id":522,"label":523,"issuer":157,"region":158,"url":524,"description":525,"useCases":526,"indexable":232},"eu-accessibility-act","European Accessibility Act","https://eur-lex.europa.eu/eli/dir/2019/882/oj","Directive (EU) 2019/882: accessibility requirements for banking services, ecommerce and other digital services, applicable since June 2025.",12,{"id":528,"label":529,"issuer":530,"region":438,"url":531,"description":532,"useCases":526,"indexable":232},"hipaa","HIPAA","US Department of Health and Human Services","https://www.hhs.gov/hipaa/index.html","US rules for the privacy and security of protected health information.",{"id":534,"label":535,"issuer":536,"region":176,"url":537,"description":538,"useCases":526,"indexable":232},"telecom-consumer-rules","Telecom consumer protection rules","National telecom regulators","https://www.berec.europa.eu/","National rules on telecom contracts, switching, billing disputes and marketing consent.",{"id":540,"label":541,"issuer":157,"region":158,"url":542,"description":543,"useCases":544,"indexable":232},"eecc","European Electronic Communications Code","https://eur-lex.europa.eu/eli/dir/2018/1972/oj","Directive (EU) 2018/1972: consumer protection, contract, switching and security rules for telecom operators.",11,{"id":546,"label":547,"issuer":548,"region":438,"url":549,"description":550,"useCases":544,"indexable":232},"us-tcpa","Telephone Consumer Protection Act","Federal Communications Commission","https://www.fcc.gov/consumers/guides/stop-unwanted-robocalls-and-texts","US consent rules for automated and prerecorded calls and texts; the FCC has confirmed AI generated voices count as artificial voices.",{"id":552,"label":553,"issuer":465,"region":271,"url":554,"description":555,"useCases":556,"indexable":232},"mas-notice-626","MAS Notice 626","https://www.mas.gov.sg/regulation/notices/notice-626","Singapore's anti money laundering and counter terrorism financing requirements for banks.",10,{"id":558,"label":559,"issuer":157,"region":158,"url":560,"description":561,"useCases":556,"indexable":232},"mifid-ii","MiFID II","https://eur-lex.europa.eu/eli/dir/2014/65/oj","Directive 2014/65/EU on markets in financial instruments: suitability and appropriateness of advice, record keeping and product governance.",{"id":563,"label":564,"issuer":157,"region":158,"url":565,"description":566,"useCases":556,"indexable":232},"eu-psd2","PSD2","https://eur-lex.europa.eu/eli/dir/2015/2366/oj","Payment Services Directive 2: strong customer authentication, transaction risk analysis exemptions and open banking access.",{"id":568,"label":569,"issuer":570,"region":158,"url":571,"description":572,"useCases":573,"indexable":232},"eba-loan-origination","EBA Guidelines on loan origination and monitoring","European Banking Authority","https://www.eba.europa.eu/regulation-and-policy/credit-risk/guidelines-on-loan-origination-and-monitoring","Expectations for credit decisioning, including the use of automated models.",9,{"id":575,"label":576,"issuer":577,"region":438,"url":578,"description":579,"useCases":580,"indexable":232},"us-ecoa-reg-b","ECOA and Regulation B","Consumer Financial Protection Bureau","https://www.consumerfinance.gov/rules-policy/regulations/1002/9/","US fair lending rules, including specific reasons in adverse action notices, which also apply when credit decisions use AI models.",8,{"id":582,"label":583,"issuer":157,"region":158,"url":584,"description":585,"useCases":580,"indexable":232},"solvency-ii","Solvency II","https://eur-lex.europa.eu/eli/dir/2009/138/oj","Directive 2009/138/EC: risk based capital, governance and model requirements for insurers.",{"id":587,"label":588,"issuer":157,"region":158,"url":589,"description":590,"useCases":363,"indexable":232},"eu-idd","Insurance Distribution Directive","https://eur-lex.europa.eu/eli/dir/2016/97/oj","Directive (EU) 2016/97: conduct rules for selling insurance, including demands and needs testing and advice.",{"id":592,"label":593,"issuer":594,"region":595,"url":596,"description":597,"useCases":337,"indexable":232},"cbuae-ai-guidance","CBUAE guidance on AI and ML","Central Bank of the UAE","middle-east","https://www.centralbank.ae/","UAE central bank expectations for the enabling technologies, AI and machine learning used by licensed financial institutions.",{"id":599,"label":600,"issuer":601,"region":158,"url":602,"description":603,"useCases":604,"indexable":232},"pra-ss1-23","PRA SS1/23 model risk management","Prudential Regulation Authority","https://www.bankofengland.co.uk/prudential-regulation/publication/2023/may/model-risk-management-principles-for-banks-ss","UK model risk management principles for banks, covering AI and machine learning models.",4,{"id":606,"label":607,"issuer":608,"region":158,"url":609,"description":610,"useCases":604,"indexable":232},"uk-psr-app-reimbursement","UK APP scam reimbursement rules","Payment Systems Regulator","https://www.psr.org.uk/our-work/app-scams/","Mandatory reimbursement of authorised push payment scam victims by UK payment firms, which shifts scam losses onto banks.",{"id":612,"label":613,"issuer":614,"region":271,"url":615,"description":616,"useCases":387,"indexable":232},"au-scams-prevention-framework","Australian Scams Prevention Framework","Australian Treasury","https://treasury.gov.au/consultation/c2024-573813","Economy wide obligations for banks, telcos and digital platforms to prevent, detect, disrupt and respond to scams.",{"id":618,"label":619,"issuer":157,"region":158,"url":620,"description":621,"useCases":387,"indexable":232},"eu-mar","EU Market Abuse Regulation","https://eur-lex.europa.eu/eli/reg/2014/596/oj","Regulation (EU) 596/2014: insider dealing and market manipulation, including the duty to detect and report suspicious orders and transactions.",{"id":623,"label":624,"issuer":625,"region":438,"url":626,"description":627,"useCases":387,"indexable":232},"us-fcra","Fair Credit Reporting Act","Federal Trade Commission","https://www.ftc.gov/legal-library/browse/statutes/fair-credit-reporting-act","US rules on consumer reports, their accuracy and permissible use, relevant to credit scoring and screening.",1790598303341]