[{"data":1,"prerenderedAt":620},["ShallowReactive",2],{"uc-public-procurement-tender-risk-screening":3,"uc-regulations":392},{"useCase":4,"evidence":186,"blitsAiDeployments":281,"benchmarks":282,"indicative":297,"related":300,"indexability":390,"includeUnpublished":192},{"title":5,"shortTitle":6,"seoTitle":7,"metaDescription":8,"status":9,"definition":10,"aliases":11,"industries":18,"functions":20,"patterns":24,"channels":28,"audience":31,"autonomy":32,"adoptionStage":33,"problem":34,"problemStats":35,"howItWorks":46,"valueDrivers":47,"kpis":51,"indicativeValue":56,"macroEstimates":79,"feasibility":80,"implementation":92,"risk":135,"blitsAi":161,"faq":163,"related":176,"datePublished":181,"dateModified":181,"lastVerified":181,"changelog":182,"slug":185},"AI screening of public tenders for fraud, rigging and errors before award","Public tender risk screening","Public procurement tender screening with AI","Brazil's CGU examined over 161,000 procurement processes in 2024 with its Alice tool, and TCU and state audit courts use it to flag risky tenders before award.","published","Public tender risk screening is AI that reads tender notices and procurement records and alerts auditors to likely errors, rigging or fraud before a contract is awarded. Audit offices and ministries use it to give auditors a short list of flagged tenders to review while a tender is still open.",[12,13,14,15,16,17],"tender risk scoring","procurement fraud detection","bid rigging detection","tender audit alerts","continuous procurement auditing","public procurement red flags",[19],"government",[21,22,23],"procurement","fraud-prevention","risk-management",[25,26,27],"document-processing","anomaly-detection","classification-and-routing",[29,30],"internal-tools","email","employee-facing","assist","early-adopters","Public purchasing is large and slow to audit. Brazil's Office of the Comptroller General (CGU)\nreports that federal purchases average 1.5 percent of GDP and that the process is perceived as\nvery vulnerable to fraud. CGU says more than 500 calls for tender are published every day, each\nwith a notice and attached documents. Reading all of them by hand is not realistic\nfor an audit team, so audit offices sample, and they often look after a contract has been signed,\nwhen the money is hard to recover.\n\nThe problems an auditor wants to catch early are specific. CGU lists tenders steered to one\ncompany, prices well above the market and purchases the organization does not need. They sit in\nthe text of the notice and in the price and quantity tables, so a tool has to read documents and\ncompare them with rules and reference prices before it can say anything useful. Alerts that\nnobody opens add to the auditors' workload and prevent nothing.",[36,41],{"statement":37,"sourceTitle":38,"sourceUrl":39,"year":40},"CGU reports that Brazilian federal government purchases correspond on average to 1.5 percent of GDP (R$ 119 billion in 2021) and that public procurement is perceived as very vulnerable to fraud.","Relato Técnico: Alice, desafios, resultados e perspectivas da ferramenta de auditoria contínua de compras públicas","https:\u002F\u002Frevista.cgu.gov.br\u002FRevista_da_CGU\u002Farticle\u002Fdownload\u002F530\u002F357\u002F3347",2022,{"statement":42,"sourceTitle":43,"sourceUrl":44,"year":45},"CGU says there are more than 500 calls for tender published every day, in its entry in the OECD Observatory of Public Sector Innovation.","Robot Alice, Bid, Contract and Notice Analyser","https:\u002F\u002Foecd-opsi.org\u002Finnovations\u002Frobot-alice-bid-contract-and-notice-analyser\u002F",2024,"1. **Collect every day.** The tool pulls new tenders, notices, attached documents, direct award\n   records and amendments from the procurement portals and the official gazette, without waiting\n   for an auditor to request them.\n2. **Read and check.** The tool reads requirements, quantities, prices, brand names and deadlines\n   with text mining, keyword patterns checked in their surrounding text and AI classification of\n   purchases, and runs them through a catalog of risk checks written by auditors. CGU described\n   about 40 such checks in 2022. A check might ask whether a specification names a\n   single brand or whether the estimated price is far above the reference.\n3. **Alert.** Each hit becomes an alert with the tender, the check that fired and the text it\n   rests on. Alerts go to the audit team that covers the buyer. CGU's 2021 agreement with the\n   Ministry of Economy was meant to let managers be alerted in the planning stage, and its June\n   2024 entry in the OECD Observatory lists that as a future step. Priority follows value,\n   criticality and relevance.\n4. **Audit while the tender is open.** For the riskiest purchases an audit is opened. The buyer\n   can then change the notice, adjust prices or quantities, or suspend or cancel the tender. CGU\n   counts these outcomes as preventive results.\n5. **Record what happened.** Auditors log whether an alert was a real finding. We recommend using\n   the hit rate of each check to decide which checks are kept, tuned or dropped. CGU's report\n   describes one such survey, in its proof of concept, after which the worst performing\n   algorithms were set aside.",[48,49,50],"risk-reduction","compliance","employee-productivity",[52,53,54,55],"cost-savings","interactions-handled","detection-rate-improvement","accuracy",{"referenceOrg":57,"inputs":58,"formula":74,"currency":75,"period":76,"resultLabel":77,"caveat":78},"A national audit office whose alerts lead to preventive audits on about EUR 1 billion of public tenders a year",[59,67],{"key":60,"label":61,"low":62,"high":63,"unit":64,"note":65,"sourceUrl":66},"auditedValue","Value of tenders put under preventive audit after an alert",500000000,2000000000,"EUR per year","Editorial assumption. For scale, CGU reported that its 2024 alerts led to audits of purchases totaling R$ 30.57 billion.","https:\u002F\u002Fwww.gov.br\u002Fcgu\u002Fpt-br\u002Fassuntos\u002Fauditoria-e-fiscalizacao\u002Falice",{"key":68,"label":69,"low":70,"high":71,"unit":72,"note":73,"sourceUrl":66},"benefitRate","Financial benefit as a share of audited value",0.01,0.04,"fraction of audited value","CGU counted R$ 1.25 billion in financial benefits in 2024 against R$ 30.57 billion in audited purchases, about 4 percent, on its own methodology. The low end is a quarter of that, an editorial assumption.","auditedValue * benefitRate","EUR","per year","Financial benefit from preventive tender audits","The benefit rate comes from the organization's own calculation and may include audits that started in an earlier year. The number leaves out the cost of building and running the tool and of the audits, and the value of tenders that were corrected without a counted saving.",[],{"complexity":81,"complexityNote":82,"dataPrerequisites":83,"integrations":88},"medium","The build needs a daily feed from the procurement portals and the gazette, a parser for notices and attachments in several formats, an audit case system that can receive alerts, and a catalog of risk checks that auditors agree on. CGU's first scraper ran into a captcha on the federal portal, which it worked around with human solving at first and later replaced with an automatic fix; a partnership with the ministry of management then widened access to reliable data. CGU names integration into the audit teams' normal work as the factor that decided success.",[84,85,86,87],"Machine readable access to notices, attached documents, bids and awards, including direct awards","Reference prices or past price data for the goods and services bought","A catalog of risk typologies with example findings from past audits","A record of audit outcomes per alert, to measure hit rates",[89,90,91],"Procurement portal and official gazette feeds","Audit case management system that can receive alerts and record outcomes","Email or workflow notifications for auditors and procurement managers",{"steps":93,"guardrails":109,"humanInTheLoop":115,"kpisToInstrument":116,"failureModes":122},[94,97,100,103,106],{"title":95,"detail":96},"Start from findings auditors already make","Turn the finding types in past audits into testable checks. CGU built its proof of concept on 13 of 21 mapped finding types, measured the hit rate of each on a sample and set aside the checks that performed worst.",{"title":98,"detail":99},"Secure the data feed first","Negotiate bulk access to notices and documents with the portal owner. CGU's first scraper hit a captcha that blocked automated downloads; human captcha solving was paid for at first, an automatic fix came later and data access widened through a partnership with the ministry of management.",{"title":101,"detail":102},"Deliver alerts inside the auditors' own system","CGU names integration into the ordinary audit workflow, with strong sponsorship from senior management, as the point that decided success, so that alerts reach auditors where they already work.",{"title":104,"detail":105},"Close the loop on every alert","Ask auditors to record the outcome. In 2021 CGU's report counted 4,528 alerts on its own jurisdiction and 646 analyses recorded in its audit system, and it notes that some issues were settled through unrecorded contact. The audit court of Minas Gerais tracked the answers to its alerts: of 60 alerts from an audit of 2,297 tenders from August 2022, 53 were answered, and notices worth R$ 91.7 million were corrected. A recorded outcome is also what lets you tune the checks.",{"title":107,"detail":108},"Offer the alerts to other audit bodies","After the core is stable, let state and municipal controllers and other branches subscribe. In 2024, 126 federal, state and municipal bodies were registered to receive CGU alerts, and the audit court of Minas Gerais reported on its own adoption.",[110,111,112,113,114],"An alert is an indication of risk; only an auditor's review turns it into a finding, and the tool never suspends or cancels a tender","Auditors check the text and data behind an alert before they contact the buyer","Every alert shows the check that fired and the document passage it rests on","Each risk check has an owner, a written definition and a measured hit rate","Personal data about named officials or sole traders is minimized and access is logged","Auditors read each alert, decide whether to open an audit and decide what to ask of the buyer. The buyer's procurement manager decides whether to change, suspend or cancel the tender.",[117,118,119,120,121],"Share of alerts analyzed by an auditor, and the time from publication to alert","Hit rate per risk check, meaning alerts confirmed as findings","Value of tenders corrected, suspended or canceled after an alert, counted separately from savings","Number of tenders covered per day against the number published","Auditor time spent per alert",[123,126,129,132],{"title":124,"detail":125},"Alerts nobody analyzes","CGU's report says Alice fell into disuse at CGU by 2018: the technology was adequate, but no standard process obliged every area to follow it. A redesign of the process, with new finding types and integration into the audit system, brought results back from 2019 to 2021. In 2021 only 646 of 4,528 alerts had a recorded analysis, though some issues may have been settled through contact that was not recorded. Limit alerts to checks with a proven hit rate, rank them and measure the share that is opened.",{"title":127,"detail":128},"Counting the wrong number as a result","CGU separates the value of tenders suspended or canceled (R$ 9.7 billion from 2019 to 2022) from the financial benefit it counted (more than R$ 1.3 billion in the same period). Quoting the first as a saving overstates the result by a large factor.",{"title":130,"detail":131},"Rules that age","Text patterns tuned to today's notices miss new drafting habits and flag legitimate specifications. Review each check against recent audit outcomes and retire checks that no longer find anything.",{"title":133,"detail":134},"Dependence on one portal","A change in the portal's structure or access rules can stop the feed overnight. Monitor feed volume daily and keep an owner for the connection.",{"euAiAct":136,"regulations":139,"guidance":144,"controls":154,"incidents":160},{"tier":137,"basis":138},"context-dependent","Screening tender documents of companies is not a use listed in Annex III, so such a system is usually not high risk. The tier changes if the system scores named individuals, such as officials or sole traders, to assess the risk that they commit a criminal offense on behalf of a law enforcement authority, which falls under Annex III point 6. GDPR applies wherever named officials, bidders who are individuals or whistleblowers appear in the data.",[140,141,142,143],"eu-ai-act","gdpr","nist-ai-rmf","iso-42001",[145,149],{"title":43,"issuer":146,"region":147,"url":44,"note":148},"OECD Observatory of Public Sector Innovation","global","Case entry submitted by CGU, with the challenges it met, the conditions for success and how the tool was shared with other bodies.",{"title":150,"issuer":151,"region":152,"url":39,"note":153},"Relato Técnico: Alice, desafios, resultados e perspectivas da ferramenta de auditoria contínua","Controladoria-Geral da União","latin-america","CGU's own technical report on how Alice was built, measured and integrated into audit work, in Portuguese.",[155,156,157,158,159],"Catalog of risk checks with owner, definition and measured hit rate","Alerts logged with the document passage and check behind them","Outcome recorded for every alert, reviewed each quarter","Separate reporting of tenders suspended and benefits counted","Access control and retention rules for personal data in tender files",[],{"howToBuild":162},"The tender data pipeline and the risk checks usually live in the audit office's analytics\nenvironment. Blits.ai adds the auditor facing layer: an **AI agent** that answers \"why was this\ntender flagged?\" from the risk engine's output through **custom functions** and a **SQL\nknowledge base**, and from the tender notice and the audit manual through a **knowledge base**\nwith hybrid retrieval. Auditors reach it in Microsoft Teams, by email or in an internal web chat.\n\nAn **agentic workflow** can assemble a case pack for each alert with the flagged clauses, reference\nprices and earlier audits on the same buyer, with **human in the loop approval** before anything\nis sent to the contracting authority. **Test suites** replay labeled tenders against the agent,\n**monitors** check its answers on a schedule and **PII masking** protects named individuals.\nRun Blits.ai with EU or UAE data residency to keep tender files in region.",[164,167,170,173],{"question":165,"answer":166},"Does AI find fraud in public tenders, or only mistakes?","The evidence on this page covers both and does not separate them. CGU says its alerts led to tenders being suspended or canceled with signs of fraud or error, and its Alice page names steering to one company, overpricing and unnecessary purchases as the problems it targets. Auditors still decide which alerts are fraud.",{"question":168,"answer":169},"How large are the reported results?","CGU examined more than 161,000 procurement processes in 2024 and counted R$ 1.25 billion in financial benefits from preventive actions, on its own calculation. Its 2022 report counted R$ 9.7 billion in tenders suspended or canceled from 2019 to June 2022. The audit court of Minas Gerais reported that an audit of 2,297 tenders from August 2022 found 60 irregularities worth R$ 113.7 million and obtained corrections of notices worth R$ 91.7 million, a value of corrected notices and not a saving. No independent audit of these figures appears in the sources.",{"question":171,"answer":172},"Can a smaller audit body use it?","CGU registered 126 federal, state and municipal bodies to receive alerts in 2024, and the audit court of Minas Gerais reported in 2023 that more than ten Brazilian courts of accounts had adopted the system. The Minas Gerais analysts named the challenges of adopting it as obtaining data, building infrastructure, selecting a typology of irregularities, designing the audit process and overcoming internal resistance.",{"question":174,"answer":175},"Is it AI or a set of rules?","It combines them. CGU's 2022 report describes robotic process automation with fixed risk checks that match keywords and read the surrounding text. It says a 2018 redesign added AI techniques to classify purchases and that further uses were being explored. Its program page now says the tool uses text mining and AI.",[177,178,179,180],"procurement-contract-review","continuous-controls-testing","inspection-prioritization","benefit-fraud-and-error-detection","2026-10-14",[183],{"date":181,"note":184},"First published","public-procurement-tender-risk-screening",[187,218,257],{"title":188,"useCases":189,"organization":190,"vendors":195,"summary":199,"stage":200,"year":201,"channels":202,"languages":203,"metrics":205,"outcomeDisclosed":192,"sources":206,"verification":213,"grade":215,"id":216,"organizationSlug":217},"Tribunal de Contas da União (Brazil): Alice analysis of tender notices",[185],{"name":191,"anonymized":192,"country":193,"region":152,"industry":19,"wikidata":194},"Tribunal de Contas da União (TCU)",false,"BR","Q7840473",[196],{"name":197,"role":198},"In house (TCU)","in-house","Brazil's federal court of accounts uses Alice (Análise de Licitações e Editais), which TCU describes as its artificial intelligence tool for analyzing procurement notices and records of electronic auctions on the federal procurement portal. Auditors told a South African delegation in September 2025 that the platform checks procurement procedures every day so the court can act before money is spent. It began as a pilot with a small team; version 3.0 added analysis of product and service descriptions and version 3.6 added more advanced automated methods to reduce inconsistencies and improve data standardization. The tool was first built by the Office of the Comptroller General and shared with TCU in 2016. The page gives no outcome figures.","production",2025,[29],[204],"pt",[],[207,212],{"url":208,"title":209,"publisher":210,"date":211},"https:\u002F\u002Fportal.tcu.gov.br\u002Fenglish\u002Fnews\u002Ftcu-presents-technological-innovations-to-south-african-delegation","TCU Presents Technological Innovations to South African Delegation","Tribunal de Contas da União","2025-09-25",{"url":44,"title":43,"publisher":146},{"level":214,"checkedAt":181},"source-verified","B","tcu-brazil-alice-procurement-analysis",null,{"title":219,"useCases":220,"organization":221,"vendors":224,"summary":227,"stage":228,"year":45,"channels":229,"languages":230,"metrics":231,"outcomeDisclosed":247,"sources":248,"verification":255,"grade":215,"id":256,"organizationSlug":217},"Controladoria-Geral da União (Brazil): Alice alerts on federal tenders",[185],{"name":222,"anonymized":192,"country":193,"region":152,"industry":19,"wikidata":223},"Controladoria-Geral da União (CGU)","Q3624988",[225],{"name":226,"role":198},"In house (CGU auditors)","Brazil's federal Office of the Comptroller General (CGU) built Alice, a tool that reads the procurement documents published every day on the federal purchasing portals, the Banco do Brasil and Caixa portals and the official gazette, applies a set of audit risk checks and sends alerts to auditors and public managers while a tender is still open. CGU describes it as using text mining and artificial intelligence; an earlier CGU technical report calls it a robotic process automation tool for continuous auditing and says that a 2018 redesign added AI techniques to classify purchases, with further uses of AI being explored. Auditors decide what to do with an alert, and CGU opens preventive audits on the highest risk purchases. In 2024 Alice examined more than 161,000 procurement processes and CGU counted R$ 1.25 billion in financial benefits from the preventive actions it directed. A 2022 CGU report says that of 4,528 alerts raised in 2021 on CGU's own jurisdiction, 646 had a recorded analysis in its audit system (managers may also have fixed issues after informal contact that was not recorded), and that R$ 9.7 billion in tenders were suspended or canceled between 2019 and June 2022. In 2024, 126 federal, state and municipal bodies were registered to receive alerts.","scaled",[29,30],[204],[232,241],{"kpi":52,"value":233,"unit":234,"currency":235,"qualifier":236,"period":237,"baseline":238,"claimant":239,"quote":240,"sourceUrl":66},1250000000,"currency","BRL","exact","2024","financial benefits counted under CGU's own calculation methodology","organization","Em 2024, foram contabilizados R$ 1,25 bilhão de benefícios financeiros decorrentes das ações preventivas direcionadas pelo Alice.",{"kpi":53,"value":242,"unit":243,"qualifier":244,"period":245,"claimant":239,"quote":246,"sourceUrl":66},161000,"count","at-least","procurement processes examined in 2024","Em 2024, o Alice examinou mais de 161 mil processos de compras, sendo 35.210 de órgãos e entidades fiscalizados pela CGU e 126.450 de outras organizações públicas.",true,[249,251,254],{"url":66,"title":250,"publisher":151},"Analisador de Licitações, Contratos e Editais (Alice)",{"url":39,"title":252,"publisher":253},"Relato Técnico: Alice, desafios, resultados e perspectivas da ferramenta de auditoria contínua de compras públicas governamentais com uso de inteligência artificial","Revista da CGU, volume 14, number 26 (July to December 2022)",{"url":44,"title":43,"publisher":146},{"level":214,"checkedAt":181},"cgu-brazil-alice-tender-alerts",{"title":258,"useCases":259,"organization":260,"vendors":263,"summary":264,"stage":200,"year":265,"channels":266,"languages":267,"metrics":268,"outcomeDisclosed":247,"sources":274,"verification":279,"grade":215,"id":280,"organizationSlug":217},"Tribunal de Contas do Estado de Minas Gerais (Brazil): Alice in tender audits",[185],{"name":261,"anonymized":192,"country":193,"region":152,"industry":19,"wikidata":262},"Tribunal de Contas do Estado de Minas Gerais (TCE MG)","Q20061562",[],"The audit court of the Brazilian state of Minas Gerais adopted Alice, the tender analysis system created by the federal Office of the Comptroller General and the federal court of accounts. On 27 June 2023 two of its analysts presented what the tool had delivered for the court. The news item reports an audit of 2,297 municipal and state tenders from August 2022, published in December. It detected 60 irregularities involving R$ 113.7 million. Of the 60 alerts sent to the audited bodies, 53 were answered in a way that allowed the control to be verified, and the audit obtained corrections of tender notices worth R$ 91.7 million. That is the value of the corrected notices, not a saving, although the headline of the news item speaks of savings. The analysts described the adoption work: obtaining data, building the infrastructure, choosing a typology of irregularities, designing the audit process and overcoming resistance inside the organization. The court said it wants to secure what has been produced, share the results and create new typologies. The news item states that more than ten Brazilian courts of accounts have adopted the system.",2023,[29],[204],[269],{"kpi":53,"value":270,"unit":243,"qualifier":236,"period":271,"claimant":239,"quote":272,"sourceUrl":273},2297,"municipal and state tenders audited from August 2022, one month of tenders","Em dezembro, o TCE publicou o resultado de uma auditoria feita em 2.297 licitações municipais e estaduais realizadas no período de um mês (agosto de 2022).","https:\u002F\u002Fwww.tce.mg.gov.br\u002Fnoticia\u002FDetalhe\u002F1111626457",[275],{"url":273,"title":276,"publisher":277,"date":278},"Tribunal mineiro apresenta economia feita com ajuda de Alice","Tribunal de Contas do Estado de Minas Gerais","2023-06-27",{"level":214,"checkedAt":181},"tce-mg-alice-procurement-audit",0,[283,291],{"kpi":53,"label":284,"unit":243,"aggregate":192,"higherIsBetter":247,"n":285,"nUpTo":281,"median":286,"min":270,"max":242,"byClaimant":287,"vendorOnly":192,"points":288},"Interactions handled",2,81648.5,{"organization":285,"vendor":281,"regulator":281,"independent":281},[289,290],{"evidenceId":256,"organization":222,"value":242,"qualifier":244,"claimant":239,"grade":215,"pooled":247},{"evidenceId":280,"organization":261,"value":270,"qualifier":236,"claimant":239,"grade":215,"pooled":247},{"kpi":52,"label":292,"unit":234,"currency":235,"aggregate":192,"higherIsBetter":247,"n":293,"nUpTo":281,"median":233,"min":233,"max":233,"byClaimant":294,"vendorOnly":192,"points":295},"Cost savings",1,{"organization":293,"vendor":281,"regulator":281,"independent":281},[296],{"evidenceId":256,"organization":222,"value":233,"qualifier":236,"claimant":239,"grade":215,"pooled":247},{"low":298,"high":299},5000000,80000000,[301,331,350,370],{"slug":177,"title":302,"shortTitle":303,"definition":304,"status":9,"industries":305,"functions":310,"patterns":313,"audience":31,"autonomy":317,"adoptionStage":33,"evidenceCount":318,"publicEvidenceCount":319,"organizations":320,"bestGrade":215,"headline":217,"lastVerified":330,"indexable":247},"AI assistant for procurement and supplier contract review","Procurement and contract review","Procurement and contract review is an assistant that reads supplier contracts and proposals, for procurement and vendor management teams. It 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.",[306,307,19,308,309],"cross-industry","banking","retail-and-ecommerce","manufacturing",[21,311,312],"legal","finance-and-accounting",[25,314,315,316],"rag-knowledge-assistant","content-generation","agentic-workflow","copilot",11,10,[321,322,323,324,325,326,327,328,329],"Bridgewater Associates","General Services Administration","Administration for Children and Families","Internal Revenue Service","Authority for Public Projects and Tenders and Local Content","Pension Benefit Guaranty Corporation","Pacific Northwest National Laboratory","Technology Common Services Center","Walmart","2026-09-27",{"slug":178,"title":332,"shortTitle":333,"definition":334,"status":9,"industries":335,"functions":338,"patterns":341,"audience":342,"autonomy":343,"adoptionStage":344,"segment":345,"evidenceCount":346,"publicEvidenceCount":346,"organizations":347,"bestGrade":215,"headline":217,"lastVerified":330,"indexable":247},"AI for continuous controls testing and control self assessment","Continuous controls testing","AI continuous controls testing checks every transaction or record against a control's rule  instead of a periodic sample, giving continuous, full population assurance. It collects evidence from source systems, maps each artifact to the control it supports, 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.",[306,307,336,337,19],"insurance","capital-markets",[23,339,340],"regulatory-compliance","operations",[316,25,26,27],"back-office","supervised-agent","emerging","second-line",3,[348,349,326],"Federal Deposit Insurance Corporation","U.S. Department of the Interior",{"slug":179,"title":351,"shortTitle":352,"definition":353,"status":9,"industries":354,"functions":355,"patterns":357,"audience":31,"autonomy":32,"adoptionStage":33,"evidenceCount":319,"publicEvidenceCount":319,"organizations":359,"bestGrade":215,"headline":217,"lastVerified":330,"indexable":247},"AI for risk based inspection prioritization in food safety, workplace and environmental regulation","Inspection prioritization","Inspection prioritization is AI that predicts which premises, operators or activities are most likely to be non compliant, to target inspectors' visits where the risk is highest. It is used across food safety, workplace safety, environmental and other regulation, ideally with inspectors still choosing visits and random inspections testing the model.",[19],[23,356,339],"case-management",[358,26],"prediction-and-scoring",[360,361,362,363,364,365,366,367,368,369],"Care Quality Commission","Driver and Vehicle Standards Agency","U.S. Environmental Protection Agency, Office of Enforcement and Compliance Assurance","U.S. Food and Drug Administration, Office of Information Operations","Food Standards Agency","Gemeente Amsterdam","Ministry of Human Resources and Emiratisation (MoHRE)","Nederlandse Arbeidsinspectie","Nederlandse Voedsel- en Warenautoriteit (NVWA)","United States Coast Guard",{"slug":180,"title":371,"shortTitle":372,"definition":373,"status":9,"industries":374,"functions":375,"patterns":377,"audience":342,"autonomy":32,"adoptionStage":33,"evidenceCount":378,"publicEvidenceCount":378,"organizations":379,"bestGrade":215,"headline":385,"lastVerified":330,"indexable":247},"AI for benefit fraud and error detection in social security","Benefit fraud and error detection","Benefit fraud and error detection uses risk models that help a social security or benefits agency decide which claims, payments and recipients to check for fraud or error. Caseworkers then 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.",[19],[22,376,356],"citizen-services",[358,26],5,[380,381,382,383,384],"Centers for Medicare and Medicaid Services","Department for Work and Pensions","Gemeente Rotterdam","U.S. Department of the Treasury, Bureau of the Fiscal Service","Uitvoeringsinstituut Werknemersverzekeringen (UWV)",{"kpi":54,"label":386,"unit":387,"n":293,"nUpTo":281,"kind":388,"value":389,"qualifier":236,"claimant":239,"organization":381,"vendorReported":192},"Detection improvement","multiplier","reported",2.5,{"indexable":247,"reasons":391},[],[393,400,405,411,418,425,431,438,446,453,459,465,471,478,485,491,497,503,509,516,521,526,532,539,545,550,555,560,567,572,580,586,592,598,604,609,614],{"id":140,"label":394,"issuer":395,"region":396,"url":397,"description":398,"useCases":399,"indexable":247},"EU AI Act","European Union","europe","https:\u002F\u002Feur-lex.europa.eu\u002Feli\u002Freg\u002F2024\u002F1689\u002Foj","Regulation (EU) 2024\u002F1689: risk based rules for AI systems, with obligations for high risk systems listed in Annex III and transparency duties under Article 50.",295,{"id":141,"label":401,"issuer":395,"region":396,"url":402,"description":403,"useCases":404,"indexable":247},"GDPR","https:\u002F\u002Feur-lex.europa.eu\u002Feli\u002Freg\u002F2016\u002F679\u002Foj","General Data Protection Regulation, including Article 22 on decisions based solely on automated processing.",265,{"id":143,"label":406,"issuer":407,"region":147,"url":408,"description":409,"useCases":410,"indexable":247},"ISO\u002FIEC 42001","ISO and IEC","https:\u002F\u002Fwww.iso.org\u002Fstandard\u002F81230.html","The international management system standard for AI.",148,{"id":142,"label":412,"issuer":413,"region":414,"url":415,"description":416,"useCases":417,"indexable":247},"NIST AI Risk Management Framework","NIST","north-america","https:\u002F\u002Fwww.nist.gov\u002Fitl\u002Fai-risk-management-framework","Voluntary US framework to map, measure, manage and govern AI risk, with a generative AI profile.",115,{"id":419,"label":420,"issuer":421,"region":396,"url":422,"description":423,"useCases":424,"indexable":247},"uk-gdpr","UK GDPR","Information Commissioner's Office","https:\u002F\u002Fico.org.uk\u002Ffor-organisations\u002Fuk-gdpr-guidance-and-resources\u002F","The UK's version of the GDPR, including rules on solely automated decisions.",84,{"id":426,"label":427,"issuer":395,"region":396,"url":428,"description":429,"useCases":430,"indexable":247},"dora","DORA","https:\u002F\u002Feur-lex.europa.eu\u002Feli\u002Freg\u002F2022\u002F2554\u002Foj","Digital Operational Resilience Act for financial entities: ICT risk, incident reporting and third party risk, including AI providers.",73,{"id":432,"label":433,"issuer":434,"region":396,"url":435,"description":436,"useCases":437,"indexable":247},"uk-consumer-duty","FCA Consumer Duty","Financial Conduct Authority","https:\u002F\u002Fwww.fca.org.uk\u002Ffirms\u002Fconsumer-duty","UK rules that require firms to deliver good outcomes for retail customers, including through automated channels.",53,{"id":439,"label":440,"issuer":441,"region":442,"url":443,"description":444,"useCases":445,"indexable":247},"mas-ai-risk-management","MAS AI risk management guidelines","Monetary Authority of Singapore","asia-pacific","https:\u002F\u002Fwww.mas.gov.sg\u002Fnews\u002Fmedia-releases\u002F2025\u002Fmas-guidelines-for-artificial-intelligence-risk-management","Singapore's supervisory expectations for AI risk management at financial institutions, building on the FEAT principles.",37,{"id":447,"label":448,"issuer":449,"region":442,"url":450,"description":451,"useCases":452,"indexable":247},"apra-cps-230","APRA CPS 230","Australian Prudential Regulation Authority","https:\u002F\u002Fwww.apra.gov.au\u002Foperational-risk-management","Australian operational risk standard covering critical operations and material service providers.",26,{"id":454,"label":455,"issuer":395,"region":396,"url":456,"description":457,"useCases":458,"indexable":247},"nis2","NIS2 Directive","https:\u002F\u002Feur-lex.europa.eu\u002Feli\u002Fdir\u002F2022\u002F2555\u002Foj","Directive (EU) 2022\u002F2555 on cybersecurity for essential and important entities, including telecom networks, energy and public administration.",23,{"id":460,"label":461,"issuer":462,"region":147,"url":463,"description":464,"useCases":458,"indexable":247},"pci-dss","PCI DSS","PCI Security Standards Council","https:\u002F\u002Fwww.pcisecuritystandards.org\u002F","Security standard for any system that stores, processes or transmits cardholder data.",{"id":466,"label":467,"issuer":468,"region":414,"url":469,"description":470,"useCases":458,"indexable":247},"us-sr-11-7","SR 11-7 model risk management","Federal Reserve and OCC","https:\u002F\u002Fwww.federalreserve.gov\u002Fboarddocs\u002Fsrletters\u002F2011\u002Fsr1107.htm","US supervisory guidance on model risk management, applied by banks to AI and machine learning models.",{"id":472,"label":473,"issuer":474,"region":414,"url":475,"description":476,"useCases":477,"indexable":247},"hipaa","HIPAA","US Department of Health and Human Services","https:\u002F\u002Fwww.hhs.gov\u002Fhipaa\u002Findex.html","US rules for the privacy and security of protected health information.",20,{"id":479,"label":480,"issuer":481,"region":396,"url":482,"description":483,"useCases":484,"indexable":247},"uk-atrs","UK Algorithmic Transparency Recording Standard","UK Government","https:\u002F\u002Fwww.gov.uk\u002Fgovernment\u002Fcollections\u002Falgorithmic-transparency-recording-standard-hub","Mandatory transparency records for algorithmic tools used by UK central government.",18,{"id":486,"label":487,"issuer":395,"region":396,"url":488,"description":489,"useCases":490,"indexable":247},"eu-accessibility-act","European Accessibility Act","https:\u002F\u002Feur-lex.europa.eu\u002Feli\u002Fdir\u002F2019\u002F882\u002Foj","Directive (EU) 2019\u002F882: accessibility requirements for banking services, ecommerce and other digital services, applicable since June 2025.",15,{"id":492,"label":493,"issuer":494,"region":147,"url":495,"description":496,"useCases":490,"indexable":247},"fatf-recommendations","FATF Recommendations","Financial Action Task Force","https:\u002F\u002Fwww.fatf-gafi.org\u002Fen\u002Fpublications\u002FFatfrecommendations\u002FFatf-recommendations.html","Global standards for anti money laundering and counter terrorist financing that national rules implement.",{"id":498,"label":499,"issuer":500,"region":414,"url":501,"description":502,"useCases":490,"indexable":247},"us-tcpa","Telephone Consumer Protection Act","Federal Communications Commission","https:\u002F\u002Fwww.fcc.gov\u002Fconsumers\u002Fguides\u002Fstop-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":504,"label":505,"issuer":395,"region":396,"url":506,"description":507,"useCases":508,"indexable":247},"eu-amlr","EU Anti Money Laundering Regulation","https:\u002F\u002Feur-lex.europa.eu\u002Feli\u002Freg\u002F2024\u002F1624\u002Foj","Regulation (EU) 2024\u002F1624: the single EU rulebook for customer due diligence, beneficial ownership and suspicious transaction reporting.",14,{"id":510,"label":511,"issuer":512,"region":414,"url":513,"description":514,"useCases":515,"indexable":247},"us-bsa","Bank Secrecy Act","FinCEN","https:\u002F\u002Fwww.fincen.gov\u002Fresources\u002Fstatutes-and-regulations\u002Fbank-secrecy-act","US anti money laundering law: customer due diligence, suspicious activity reports and record keeping.",13,{"id":517,"label":518,"issuer":395,"region":396,"url":519,"description":520,"useCases":515,"indexable":247},"eu-psd2","PSD2","https:\u002F\u002Feur-lex.europa.eu\u002Feli\u002Fdir\u002F2015\u002F2366\u002Foj","Payment Services Directive 2: strong customer authentication, transaction risk analysis exemptions and open banking access.",{"id":522,"label":523,"issuer":395,"region":396,"url":524,"description":525,"useCases":515,"indexable":247},"solvency-ii","Solvency II","https:\u002F\u002Feur-lex.europa.eu\u002Feli\u002Fdir\u002F2009\u002F138\u002Foj","Directive 2009\u002F138\u002FEC: risk based capital, governance and model requirements for insurers.",{"id":527,"label":528,"issuer":529,"region":147,"url":530,"description":531,"useCases":515,"indexable":247},"telecom-consumer-rules","Telecom consumer protection rules","National telecom regulators","https:\u002F\u002Fwww.berec.europa.eu\u002F","National rules on telecom contracts, switching, billing disputes and marketing consent.",{"id":533,"label":534,"issuer":535,"region":414,"url":536,"description":537,"useCases":538,"indexable":247},"us-ecoa-reg-b","ECOA and Regulation B","Consumer Financial Protection Bureau","https:\u002F\u002Fwww.consumerfinance.gov\u002Frules-policy\u002Fregulations\u002F1002\u002F9\u002F","US fair lending rules, including specific reasons in adverse action notices, which also apply when credit decisions use AI models.",12,{"id":540,"label":541,"issuer":542,"region":396,"url":543,"description":544,"useCases":318,"indexable":247},"eba-loan-origination","EBA Guidelines on loan origination and monitoring","European Banking Authority","https:\u002F\u002Fwww.eba.europa.eu\u002Fregulation-and-policy\u002Fcredit-risk\u002Fguidelines-on-loan-origination-and-monitoring","Expectations for credit decisioning, including the use of automated models.",{"id":546,"label":547,"issuer":395,"region":396,"url":548,"description":549,"useCases":318,"indexable":247},"eecc","European Electronic Communications Code","https:\u002F\u002Feur-lex.europa.eu\u002Feli\u002Fdir\u002F2018\u002F1972\u002Foj","Directive (EU) 2018\u002F1972: consumer protection, contract, switching and security rules for telecom operators.",{"id":551,"label":552,"issuer":395,"region":396,"url":553,"description":554,"useCases":318,"indexable":247},"mifid-ii","MiFID II","https:\u002F\u002Feur-lex.europa.eu\u002Feli\u002Fdir\u002F2014\u002F65\u002Foj","Directive 2014\u002F65\u002FEU on markets in financial instruments: suitability and appropriateness of advice, record keeping and product governance.",{"id":556,"label":557,"issuer":441,"region":442,"url":558,"description":559,"useCases":319,"indexable":247},"mas-notice-626","MAS Notice 626","https:\u002F\u002Fwww.mas.gov.sg\u002Fregulation\u002Fnotices\u002Fnotice-626","Singapore's anti money laundering and counter terrorism financing requirements for banks.",{"id":561,"label":562,"issuer":563,"region":414,"url":564,"description":565,"useCases":566,"indexable":247},"us-fcra","Fair Credit Reporting Act","Federal Trade Commission","https:\u002F\u002Fwww.ftc.gov\u002Flegal-library\u002Fbrowse\u002Fstatutes\u002Ffair-credit-reporting-act","US rules on consumer reports, their accuracy and permissible use, relevant to credit scoring and screening.",8,{"id":568,"label":569,"issuer":395,"region":396,"url":570,"description":571,"useCases":566,"indexable":247},"eu-idd","Insurance Distribution Directive","https:\u002F\u002Feur-lex.europa.eu\u002Feli\u002Fdir\u002F2016\u002F97\u002Foj","Directive (EU) 2016\u002F97: conduct rules for selling insurance, including demands and needs testing and advice.",{"id":573,"label":574,"issuer":575,"region":576,"url":577,"description":578,"useCases":579,"indexable":247},"cbuae-ai-guidance","CBUAE guidance on AI and ML","Central Bank of the UAE","middle-east","https:\u002F\u002Fwww.centralbank.ae\u002F","UAE central bank expectations for the enabling technologies, AI and machine learning used by licensed financial institutions.",6,{"id":581,"label":582,"issuer":583,"region":396,"url":584,"description":585,"useCases":378,"indexable":247},"uk-psr-app-reimbursement","UK APP scam reimbursement rules","Payment Systems Regulator","https:\u002F\u002Fwww.psr.org.uk\u002Four-work\u002Fapp-scams\u002F","Mandatory reimbursement of authorised push payment scam victims by UK payment firms, which shifts scam losses onto banks.",{"id":587,"label":588,"issuer":395,"region":396,"url":589,"description":590,"useCases":591,"indexable":247},"eu-mortgage-credit-directive","EU Mortgage Credit Directive","https:\u002F\u002Feur-lex.europa.eu\u002Feli\u002Fdir\u002F2014\u002F17\u002Foj","Directive 2014\u002F17\u002FEU: creditworthiness assessment, disclosure and advice rules for residential mortgage lending.",4,{"id":593,"label":594,"issuer":595,"region":396,"url":596,"description":597,"useCases":591,"indexable":247},"pra-ss1-23","PRA SS1\u002F23 model risk management","Prudential Regulation Authority","https:\u002F\u002Fwww.bankofengland.co.uk\u002Fprudential-regulation\u002Fpublication\u002F2023\u002Fmay\u002Fmodel-risk-management-principles-for-banks-ss","UK model risk management principles for banks, covering AI and machine learning models.",{"id":599,"label":600,"issuer":601,"region":442,"url":602,"description":603,"useCases":346,"indexable":247},"au-scams-prevention-framework","Australian Scams Prevention Framework","Australian Treasury","https:\u002F\u002Ftreasury.gov.au\u002Fconsultation\u002Fc2024-573813","Economy wide obligations for banks, telcos and digital platforms to prevent, detect, disrupt and respond to scams.",{"id":605,"label":606,"issuer":395,"region":396,"url":607,"description":608,"useCases":346,"indexable":247},"eu-mar","EU Market Abuse Regulation","https:\u002F\u002Feur-lex.europa.eu\u002Feli\u002Freg\u002F2014\u002F596\u002Foj","Regulation (EU) 596\u002F2014: insider dealing and market manipulation, including the duty to detect and report suspicious orders and transactions.",{"id":610,"label":611,"issuer":441,"region":442,"url":612,"description":613,"useCases":346,"indexable":247},"mas-shared-responsibility-framework","MAS Shared Responsibility Framework","https:\u002F\u002Fwww.mas.gov.sg\u002Fregulation\u002Fguidelines\u002Fguidelines-on-shared-responsibility-framework","Singapore's framework that assigns banks and telcos duties and loss sharing for phishing scams.",{"id":615,"label":616,"issuer":617,"region":414,"url":618,"description":619,"useCases":346,"indexable":247},"nyc-local-law-144","NYC Local Law 144","New York City","https:\u002F\u002Fwww.nyc.gov\u002Fsite\u002Fdca\u002Fabout\u002Fautomated-employment-decision-tools.page","Bias audits and notices for automated employment decision tools used in hiring and promotion in New York City.",1792049864756]