[{"data":1,"prerenderedAt":657},["ShallowReactive",2],{"uc-ediscovery-and-disclosure-document-review":3,"uc-regulations":449},{"useCase":4,"evidence":208,"blitsAiDeployments":341,"benchmarks":342,"indicative":367,"related":370,"indexability":447,"includeUnpublished":214},{"title":5,"shortTitle":6,"seoTitle":5,"metaDescription":7,"status":8,"definition":9,"aliases":10,"industries":18,"functions":22,"patterns":25,"channels":29,"audience":31,"autonomy":32,"adoptionStage":33,"problem":34,"problemStats":35,"howItWorks":43,"valueDrivers":44,"kpis":50,"indicativeValue":55,"macroEstimates":90,"feasibility":91,"implementation":104,"risk":147,"blitsAi":184,"faq":186,"related":199,"datePublished":203,"dateModified":203,"lastVerified":203,"changelog":204,"slug":207},"AI for eDiscovery and disclosure document review","eDiscovery document review","AI ranks and codes millions of documents for relevance and privilege so lawyers review fewer. The UK Serious Fraud Office and the US FTC use it in investigations.","published","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.",[11,12,13,14,15,16,17],"technology assisted review","TAR","predictive coding","AI document review","generative AI eDiscovery","privilege review AI","AI disclosure review",[19,20,21],"cross-industry","professional-services","government",[23,24],"legal","case-management",[26,27,28],"classification-and-routing","document-processing","summarization",[30],"internal-tools","employee-facing","copilot","mainstream","Every lawsuit, investigation and regulatory information request starts with a collection of\nelectronic material that nobody can read in full. A single civil matter can involve more than\n300,000 documents, and some Serious Fraud Office cases start with up to 60 million. Lawyers\nstill have to find what is relevant, what supports or undermines each side, and what is\nprivileged, under deadlines set by a court or a regulator.\n\nKeyword searches were the first answer, and they are weak: they miss documents that use other\nwords, return large volumes of noise, and give a false sense of completeness. Linear review by\ncontract lawyers is slow and expensive, and the US Department of Justice lists errors and speed\ndelays in exclusively manual review of voluminous electronic information as the problem its\neLitigation tools address. In criminal\ncases the stakes are higher still. The UK Serious Fraud Office offered no evidence in its G4S case\nafter ten years, with disclosure featuring as a core reason, and its inspectorate found that a\nmisunderstanding of how searches worked in its older review system compounded the problems.",[36,41],{"statement":37,"sourceTitle":38,"sourceUrl":39,"year":40},"HM Crown Prosecution Service Inspectorate reports that the Serious Fraud Office roughly estimates that managing and handling disclosure takes 25% of its operational budget and 40% of its staff capacity.","Serious Fraud Office: Disclosure. An inspection of the handling and management of disclosure in the Serious Fraud Office","https://cdn.websitebuilder.service.justice.gov.uk/uploads/sites/24/2024/08/SFO-Disclosure-Report-2.pdf",2024,{"statement":42,"sourceTitle":38,"sourceUrl":39,"year":40},"HM Crown Prosecution Service Inspectorate reports that some Serious Fraud Office cases start with up to 60 million documents.","1. **Collect and process.** Material from mailboxes, devices, chat tools and file shares is\n   loaded into a review platform, text is extracted, duplicates and near duplicates are removed\n   and email threads are grouped.\n2. **Explore early.** Clustering, timelines and concept search show what the collection is\n   about, so the team can exclude irrelevant data and agree search parameters before paid review\n   starts.\n3. **Rank or code.** Either a classifier learns from reviewers' decisions and keeps ranking the\n   remaining documents by likely relevance (active learning), or a large language model reads each\n   document against written review instructions and returns a relevance call, the issues it\n   touches and a short rationale.\n4. **Screen for privilege and sensitivity.** A separate pass flags likely privileged documents,\n   personal data and material that needs redaction, for lawyers to confirm.\n5. **Validate.** Random samples from the documents classed as relevant and not relevant are\n   reviewed by people to estimate precision, recall and elusion, and the method and results are\n   recorded so they can be explained to the other side or the court.\n6. **Review and produce.** Lawyers review the prioritised set, decide what to disclose, withhold\n   or redact, and the platform keeps the audit trail.",[45,46,47,48,49],"cost-to-serve","speed","employee-productivity","compliance","risk-reduction",[51,52,53,54],"processing-time-reduction","hours-saved","cost-savings","interactions-handled",{"referenceOrg":56,"inputs":57,"formula":85,"currency":86,"period":87,"resultLabel":88,"caveat":89},"A litigation or investigations team that reviews 1 million documents a year",[58,64,71,78],{"key":59,"label":60,"low":61,"high":61,"unit":62,"note":63},"documents","Documents collected for review per year after deduplication",1000000,"documents per year","The reference team. Replace with your own review volumes.",{"key":65,"label":66,"low":67,"high":68,"unit":69,"note":70},"hoursPerThousand","Reviewer hours per 1,000 documents in a linear first pass review",15,25,"hours per 1,000 documents","Editorial assumption, replace with your own review rates.",{"key":72,"label":73,"low":74,"high":75,"unit":76,"note":77},"reviewAvoided","Share of first pass review hours avoided by prioritisation or AI coding",0.3,0.6,"fraction of review hours","Conservative against the vendor reported 85% reduction in review time for one Purpose Legal matter on this page, because validation sampling, privilege review and quality control still need people.",{"key":79,"label":80,"low":81,"high":82,"unit":83,"note":84},"hourlyCost","Fully loaded cost of a reviewer hour",50,100,"USD per hour","Editorial assumption covering contract reviewers and supervising associates.","documents / 1000 * hoursPerThousand * reviewAvoided * hourlyCost","USD","per year","First pass review cost avoided","Covers first pass review labour only. It leaves out platform and model costs, processing and hosting fees, the cost of validation and of any dispute about the method, and the value of meeting deadlines that manual review would miss.",[],{"complexity":92,"complexityNote":93,"dataPrerequisites":94,"integrations":99},"medium","The technology is mature and widely available inside review platforms. The work is in the review protocol, the validation method, privilege handling and agreeing the approach with the other side, the court or the regulator, and in training case teams to use it properly.",[95,96,97,98],"Processed collections with extracted text, metadata and deduplication","Written review instructions per issue, with examples of relevant and not relevant documents","A seed or control set coded by lawyers who know the case","A sampling plan for validation with target recall agreed in advance",[100,101,102,103],"eDiscovery or review platform (processing, review, production)","Legal hold and collection tools for mailboxes, devices and chat","Matter management and privilege log tooling","Secure hosting in the jurisdiction the data must stay in",{"steps":105,"guardrails":121,"humanInTheLoop":127,"kpisToInstrument":128,"failureModes":134},[106,109,112,115,118],{"title":107,"detail":108},"Agree the protocol before the model","Write the review questions, issue definitions and privilege rules first, and decide how success will be measured (recall target, sample sizes). Where the other side or a court will scrutinise the process, share the approach early rather than defend it later.",{"title":110,"detail":111},"Pilot on a coded sample","Run the classifier or the language model on a few hundred documents that lawyers have already coded, compare, and refine the instructions until the disagreements are understood. Relativity reports that Purpose Legal reached a workable prompt for ten issues after three iterations on a sample of fewer than 500 documents.",{"title":113,"detail":114},"Run, rank and route","Apply the model to the full population, send the highest ranked documents to human review first, and keep a separate privilege and personal data pass.",{"title":116,"detail":117},"Validate with statistics, not impressions","Draw random samples from both the relevant and the not relevant sets, have people review them blind, estimate recall and elusion, and document every step so it can be explained.",{"title":119,"detail":120},"Train the case team","Make sure the people running the review understand what the tool does and does not do. The inspectorate warned of a risk that some SFO staff are not confident using the new platform, found that a number were not using it to its full potential and that many saw the training as inadequate, and found that misunderstandings about search had contributed to earlier failures.",[122,123,124,125,126],"No document is withheld as privileged or produced without a lawyer's decision","Validation sampling with a recall estimate before any review is declared complete","Written record of search terms, model instructions, versions and sampling results","Separate handling of privileged and personal data, with redaction checked by a person","Data stays in the hosting region and is not used to train a vendor's general models","Lawyers write the review instructions, code the training or validation samples, decide every privilege call and every production, and sign off the validation results. The AI decides only the order and the first view of relevance.",[129,130,131,132,133],"Estimated recall and elusion rate from validation samples, per matter","Reviewer hours per 1,000 documents, before and after","Share of the collection reviewed by people","Days from collection to production","Privilege clawback requests and disclosure errors found later",[135,138,141,144],{"title":136,"detail":137},"False confidence in completeness","Teams treat search or model output as if it found everything. HMCPSI warned that searching millions of documents is not an exact science. Measure recall and say what it is.",{"title":139,"detail":140},"Processing gaps upstream","Documents that were never extracted or indexed properly cannot be found by any model. Check processing exceptions, container files and encoding before trusting results.",{"title":142,"detail":143},"Instructions that drift from the case","The issues change as the case develops but the model instructions do not. Version the instructions and rerun validation when they change.",{"title":145,"detail":146},"Unexplainable method","The team cannot describe how documents were excluded when challenged. Keep the protocol, versions and samples as part of the matter record.",{"euAiAct":148,"regulations":151,"guidance":157,"controls":173,"incidents":179},{"tier":149,"basis":150},"context-dependent","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.",[152,153,154,155,156],"eu-ai-act","gdpr","uk-gdpr","iso-42001","nist-ai-rmf",[158,164,169],{"title":159,"issuer":160,"region":161,"url":162,"note":163},"Attorney General's Guidelines on Disclosure (2024)","UK Attorney General's Office","europe","https://www.gov.uk/government/publications/attorney-generals-guidelines-on-disclosure","Where examining every item of seized material would be disproportionate, disclosure officers can apply search techniques under Annex A, and must record the reasons for their approach in writing.",{"title":165,"issuer":166,"region":161,"url":167,"note":168},"Disclosure Review Working Group considering simplification of Practice Direction 57AD","Courts and Tribunals Judiciary (England and Wales)","https://www.judiciary.uk/disclosure-review-working-group-considering-simplification-of-practice-direction-57ad/","A judiciary led group is reviewing civil disclosure rules in the Business and Property Courts, including the use of technology assisted review and AI.",{"title":170,"issuer":171,"region":161,"url":39,"note":172},"Serious Fraud Office: Disclosure, an inspection report","HM Crown Prosecution Service Inspectorate","Describes how a prosecutor uses an AI enabled review platform, the limits of search, and the need to train staff and quality assure review.",[174,175,176,177,178],"Documented review protocol and validation plan for every matter","Named lawyer accountable for the review method and its explanation","Audit trail of model versions, instructions, coding decisions and samples","Data processing agreements and hosting location that match the data's origin","Quality assurance batches reviewed by a second person",[180],{"title":181,"url":182,"note":183},"SFO drops decade-long probe and prosecution and announces further issue with legacy disclosure system","https://www.lawgazette.co.uk/news/sfo-investigates-troubling-disclosure-issues/5125889.article","The Law Gazette reports that the Serious Fraud Office found a further problem in its legacy Autonomy review system, in how some digital container files were expanded, so that some items may not have been available for review in about 20 long running cases; an earlier review covered 66 historical conviction cases. Not an AI model error, but a reminder that review results are only as complete as the processing underneath them.",{"howToBuild":185},"Blits.ai is not a review platform, so the review itself stays in the organization's eDiscovery\ntool. What can be built on Blits.ai is the layer around it: an **agentic workflow** that takes a\nbatch of extracted document text through the **REST API**, asks an **agent** with a versioned\nprompt and **structured output** for a relevance call, the issues touched and a rationale per\ndocument, and writes the result back through **custom functions**, with a full **audit trail**\nper run and **human in the loop** approval before any bulk coding is applied.\n\n**PII masking** at the gateway limits the personal data the model sees, **EU or UAE data\nresidency** keeps the data in region, and the platform is **model agnostic**, so the team can compare models on the same instructions.\n**Test suites** with lawyer coded documents act as a regression set when the instructions change,\nand a **knowledge base** with the review protocol and issue definitions lets reviewers ask\nquestions about the matter's rules while they work.",[187,190,193,196],{"question":188,"answer":189},"Can we rely on AI review when the other side or the court will scrutinise it?","Only with a method you can explain. In England and Wales the Attorney General's disclosure guidelines let investigators use search techniques when reviewing everything would be disproportionate, provided they record their reasons, and the judiciary is reviewing civil disclosure rules with technology assisted review and AI in mind. Agreed instructions, statistical validation and a written record are what make the result defensible.",{"question":191,"answer":192},"How much review time does it save?","It depends on the collection and the recall target. Relativity reports that Purpose Legal cut project time by 85% (4,000 review hours) on one 300,000 document matter, measured against a contract review that would have taken multiple weeks and was never run. The UK Serious Fraud Office used its pilot tool to screen about 30 million documents for privilege in the Rolls-Royce case, work that independent barristers had done by hand. Plan on people still reviewing the prioritised set, the privilege calls and the validation samples.",{"question":194,"answer":195},"What is the difference between active learning and generative AI review?","Active learning trains a classifier on reviewers' decisions and keeps reranking the rest; the FTC uses Relativity Active Learning to predict pertinent documents. Generative AI review has a language model read each document against written instructions and explain its call. Both need the same validation.",{"question":197,"answer":198},"Is AI document review high risk under the EU AI Act?","For a company or law firm reviewing documents in civil litigation, usually not. It is high risk when law enforcement uses it to evaluate the reliability of evidence in criminal cases, or when a judicial authority uses it to research and apply the law.",[200,201,202],"court-and-case-file-summarization","freedom-of-information-request-processing","intelligent-document-processing","2026-09-27",[205],{"date":203,"note":206},"First published","ediscovery-and-disclosure-document-review",[209,243,262,298],{"title":210,"useCases":211,"organization":212,"vendors":217,"summary":227,"stage":228,"year":40,"channels":229,"languages":230,"metrics":232,"outcomeDisclosed":214,"sources":233,"verification":238,"grade":240,"id":241,"organizationSlug":242},"US Department of Justice: AI features in eLitigation tools for review of voluminous electronic evidence",[207],{"name":213,"anonymized":214,"country":215,"region":216,"industry":21},"U.S. Department of Justice",false,"US","north-america",[218,221,223,225],{"name":219,"role":220},"Relativity","platform",{"name":222,"role":220},"Everlaw",{"name":224,"role":220},"Nuix",{"name":226,"role":220},"CloudNine Law","Department of Justice components use commercial eLitigation platforms such as Relativity and Everlaw for investigations, litigation and FOIA and Privacy Act work, and these tools increasingly include AI. The department's inventory entry describes surfacing potentially discoverable material in large collections of emails, text messages and other records, locating potentially inculpatory or exculpatory evidence, and identifying material to disclose or withhold under legal rules and privileges. The entry is department wide and deployed since January 2024. It is formally classified as \"Presumed High-Impact, but Not High-impact\", because the output is not the principal basis for decisions with legal or similarly significant effect, while noting that the tools are used in high impact contexts and that individual uses vary. No outcome figures are published.","scaled",[30],[231],"en",[],[234],{"url":235,"title":236,"publisher":237},"https://raw.githubusercontent.com/ombegov/2025-Federal-Agency-AI-Use-Case-Inventory/main/Data/2025_individually_reported_AI_use_cases.csv","2025 individually reported AI use cases (entry DOJ-0102, eLitigation Tools)","Office of Management and Budget (GitHub)",{"level":239,"checkedAt":203},"source-verified","B","doj-elitigation-ai-document-review","u-s-department-of-justice",{"title":244,"useCases":245,"organization":246,"vendors":248,"summary":250,"stage":251,"year":252,"channels":253,"languages":254,"metrics":255,"outcomeDisclosed":214,"sources":256,"verification":259,"grade":240,"id":260,"organizationSlug":261},"Federal Trade Commission: active learning to prioritise documents in investigations and litigation",[207],{"name":247,"anonymized":214,"country":215,"region":216,"industry":21},"Federal Trade Commission",[249],{"name":219,"role":220},"The Federal Trade Commission's Bureau of Competition, Bureau of Consumer Protection and Office of the General Counsel use Relativity Active Learning in eDiscovery for consumer protection and competition investigations and litigation. The inventory entry gives the problem as manual document coding being very time consuming during legal review and the output as predicted pertinent documents. It lists the use as deployed since 2020 and not high impact. No outcome figures are published.","production",2020,[30],[231],[],[257],{"url":235,"title":258,"publisher":237},"2025 individually reported AI use cases (entry FTC-0008, Relativity Active Learning)",{"level":239,"checkedAt":203},"ftc-relativity-active-learning-review",null,{"title":263,"useCases":264,"organization":265,"vendors":268,"summary":271,"stage":228,"year":272,"channels":273,"languages":274,"metrics":275,"outcomeDisclosed":284,"sources":285,"verification":296,"grade":240,"id":297,"organizationSlug":261},"Serious Fraud Office: AI document review for privilege and disclosure in criminal cases",[207],{"name":266,"anonymized":214,"country":267,"region":161,"industry":21},"Serious Fraud Office","GB",[269],{"name":270,"role":220},"OpenText","The UK Serious Fraud Office first used an AI tool in its Rolls-Royce investigation to screen about 30 million documents for material potentially covered by legal professional privilege, work that independent barristers had previously done by hand. From April 2018 it made AI document review available to all new cases and adopted OpenText Axcelerate, which groups documents by subject, builds timelines and removes duplicates; at launch the SFO said it would eventually be able to sift for relevancy. The 2024 inspection by HM Crown Prosecution Service Inspectorate describes Axcelerate's artificial intelligence and machine learning features and how document reviewers tag material for relevancy in it. It warned that there remains a risk that some staff are not confident using the system, found that a number of staff were not using it to its full potential and that many viewed its training as inadequate, and warned that searching millions of documents is not an exact science. It also found that disclosure problems in an earlier case were compounded by a misunderstanding of how searches worked in the SFO's previous document review system, which is no longer used.",2018,[30],[231],[276],{"kpi":54,"value":277,"unit":278,"qualifier":279,"period":280,"claimant":281,"quote":282,"sourceUrl":283},30000000,"count","approximately","Rolls-Royce investigation pilot, privilege screening","organization","It enabled the estimated 30 million documents provided by the company to be analysed for material potentially covered by Legal Professional Privilege.","https://www.wired-gov.net/wg/news.nsf/articles/AI+powered+RoboLawyer+helps+step+up+the+SFOs+fight+against+economic+crime+11042018162000?open=",true,[286,290,295],{"url":283,"title":287,"publisher":288,"date":289},"AI powered 'Robo-Lawyer' helps step up the SFO's fight against economic crime (official press release)","Serious Fraud Office, republished by Wired-Gov","2018-04-11",{"url":291,"title":292,"publisher":266,"date":293,"archivedUrl":294},"https://www.sfo.gov.uk/2018/04/10/ai-powered-robo-lawyer-helps-step-up-the-sfos-fight-against-economic-crime/","AI powered 'Robo-Lawyer' helps step up the SFO's fight against economic crime","2018-04-10","https://web.archive.org/web/20240225110751/https://www.sfo.gov.uk/2018/04/10/ai-powered-robo-lawyer-helps-step-up-the-sfos-fight-against-economic-crime/",{"url":39,"title":38,"publisher":171},{"level":239,"checkedAt":203},"serious-fraud-office-ai-document-review",{"title":299,"useCases":300,"organization":301,"vendors":303,"summary":305,"stage":251,"year":306,"channels":307,"languages":308,"metrics":309,"outcomeDisclosed":284,"sources":334,"verification":338,"grade":339,"id":340,"organizationSlug":261},"Purpose Legal: generative AI issues review of 300,000 documents for a court deadline",[207],{"name":302,"anonymized":214,"country":215,"region":216,"industry":20},"Purpose Legal",[304],{"name":219,"role":220},"A law firm that took over a matter as new counsel had one week to review more than 300,000 documents for a court ordered production, with each produced document mapped to the requests for production. Its eDiscovery provider Purpose Legal ran Relativity aiR for Review, a large language model review tool, refining the prompt for ten issues with a firm partner on a sample of fewer than 500 documents before running the full set. The team validated the result with random precision and elusion samples and reports recall above 95%. The results are published by the vendor, and the time and cost savings are estimated against a traditional Active Learning or TAR 2.0 contract review that would have taken multiple weeks and was never run.",2025,[30],[231],[310,318,324,330],{"kpi":51,"value":311,"unit":312,"qualifier":313,"baseline":314,"claimant":315,"quote":316,"sourceUrl":317},85,"percent","exact","project time compared with a traditional Active Learning or TAR 2.0 contract review that the team says would have taken multiple weeks; that review never ran, so this is an estimate against a counterfactual, not a measured comparison","vendor","Used aiR for Review’s issues analysis to identify documents responsive to 10 key issues – resulting in an 85% reduction in project time.","https://www.relativity.com/resources/customers/purpose-legal-relativity-air/",{"kpi":52,"value":319,"unit":320,"qualifier":313,"period":321,"baseline":322,"claimant":315,"quote":323,"sourceUrl":317},4000,"hours","one matter","reviewer hours compared with the same counterfactual Active Learning or TAR 2.0 contract review","Faced with an aggressive deadline and limited budget, aiR for Review allowed the firm to easily complete the review with a skeleton team – reducing review time by 85%, or 4,000 hours.",{"kpi":53,"value":325,"unit":326,"currency":86,"qualifier":327,"period":321,"baseline":328,"claimant":315,"quote":329,"sourceUrl":317},70000,"currency","at-least","compared with the same counterfactual contract review","This resulted in cost savings of over $70,000 for the law firm, who were thrilled with how aiR for Review let them breathe easy in a tough spot.",{"kpi":54,"value":331,"unit":278,"qualifier":327,"period":332,"claimant":315,"quote":333,"sourceUrl":317},300000,"seven days, one matter","Over 300,000 documents were reviewed in just seven days, using only one project manager and a law firm partner to provide subject matter expertise.",[335],{"url":317,"title":336,"publisher":219,"archivedUrl":337},"Purpose Legal Slashes Thousands of Hours to Beat Impossible Deadline with Relativity aiR for Review","https://web.archive.org/web/20251006015118/https://www.relativity.com/resources/customers/purpose-legal-relativity-air/",{"level":239,"checkedAt":203},"C","purpose-legal-generative-ai-issues-review",0,[343,352,357,362],{"kpi":54,"label":344,"unit":278,"aggregate":214,"higherIsBetter":284,"n":345,"nUpTo":341,"median":346,"min":331,"max":277,"byClaimant":347,"vendorOnly":214,"points":349},"Interactions handled",2,15150000,{"organization":348,"vendor":348,"regulator":341,"independent":341},1,[350,351],{"evidenceId":297,"organization":266,"value":277,"qualifier":279,"claimant":281,"grade":240,"pooled":284},{"evidenceId":340,"organization":302,"value":331,"qualifier":327,"claimant":315,"grade":339,"pooled":284},{"kpi":53,"label":353,"unit":326,"currency":86,"aggregate":214,"higherIsBetter":284,"n":348,"nUpTo":341,"median":325,"min":325,"max":325,"byClaimant":354,"vendorOnly":284,"points":355},"Cost savings",{"organization":341,"vendor":348,"regulator":341,"independent":341},[356],{"evidenceId":340,"organization":302,"value":325,"qualifier":327,"claimant":315,"grade":339,"pooled":284},{"kpi":51,"label":358,"unit":312,"aggregate":284,"higherIsBetter":284,"n":348,"nUpTo":341,"median":311,"min":311,"max":311,"byClaimant":359,"vendorOnly":284,"points":360},"Cycle time reduction",{"organization":341,"vendor":348,"regulator":341,"independent":341},[361],{"evidenceId":340,"organization":302,"value":311,"qualifier":313,"claimant":315,"grade":339,"pooled":284},{"kpi":52,"label":363,"unit":320,"aggregate":214,"higherIsBetter":284,"n":348,"nUpTo":341,"median":319,"min":319,"max":319,"byClaimant":364,"vendorOnly":284,"points":365},"Hours saved",{"organization":341,"vendor":348,"regulator":341,"independent":341},[366],{"evidenceId":340,"organization":302,"value":319,"qualifier":313,"claimant":315,"grade":339,"pooled":284},{"low":368,"high":369},225000,1500000,[371,385,399,426],{"slug":200,"title":372,"shortTitle":373,"definition":374,"status":8,"industries":375,"functions":376,"patterns":377,"audience":31,"autonomy":32,"adoptionStage":379,"evidenceCount":380,"publicEvidenceCount":380,"organizations":381,"bestGrade":240,"headline":261,"lastVerified":203,"indexable":284},"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.",[21],[23,24],[28,27,378],"rag-knowledge-assistant","early-adopters",4,[382,213,383,384],"Crown Prosecution Service","Gemeente Amsterdam","Supremo Tribunal Federal",{"slug":201,"title":386,"shortTitle":387,"definition":388,"status":8,"industries":389,"functions":390,"patterns":392,"audience":31,"autonomy":32,"adoptionStage":379,"evidenceCount":380,"publicEvidenceCount":380,"organizations":394,"bestGrade":240,"headline":261,"lastVerified":398,"indexable":284},"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.",[21],[391,23,24],"citizen-services",[27,26,393],"content-generation",[213,395,396,397],"U.S. Food and Drug Administration, Center for Drug Evaluation and Research","Provincie Noord-Holland","U.S. Department of the Interior","2026-09-26",{"slug":202,"title":400,"shortTitle":401,"definition":402,"status":8,"industries":403,"functions":406,"patterns":409,"audience":411,"autonomy":412,"adoptionStage":33,"evidenceCount":413,"publicEvidenceCount":414,"organizations":415,"bestGrade":240,"headline":421,"lastVerified":203,"indexable":284},"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.",[19,21,404,405],"automotive","manufacturing",[407,24,408],"operations","finance-and-accounting",[27,410,26],"computer-vision","back-office","supervised-agent",7,5,[416,417,418,419,420],"Ancine","Pupuk Indonesia","U.S. Immigration and Customs Enforcement","U.S. Citizenship and Immigration Services","Volvo Group",{"kpi":422,"label":423,"unit":312,"n":348,"nUpTo":341,"kind":424,"value":425,"qualifier":327,"claimant":315,"organization":416,"vendorReported":284},"accuracy","Accuracy","reported",90,{"slug":427,"title":428,"shortTitle":429,"definition":430,"status":8,"industries":431,"functions":434,"patterns":436,"audience":411,"autonomy":412,"adoptionStage":33,"segment":411,"evidenceCount":437,"publicEvidenceCount":437,"organizations":438,"bestGrade":240,"headline":445,"lastVerified":203,"indexable":284},"correspondence-triage-and-routing","AI for inbound correspondence triage and routing","Correspondence triage and routing","AI that sorts inbound correspondence before anyone answers it: it takes every inbound letter, email, upload and secure message into one intake, identifies what it is, extracts the key fields, links it to the right customer and account, sets priority and routes it to the right team or workflow, replacing the manual sorting desk.",[19,432,433,21],"banking","insurance",[407,435,24],"customer-service",[26,27,28],6,[439,440,441,442,443,444],"Ecclesia Group","Encova Insurance","Loadsure","The Master Trust Bank of Japan","Travelers","U.S. Department of Veterans Affairs",{"kpi":422,"label":423,"unit":312,"n":348,"nUpTo":341,"kind":424,"value":446,"qualifier":313,"claimant":315,"organization":443,"vendorReported":284},91,{"indexable":284,"reasons":448},[],[450,456,461,468,474,480,486,493,501,507,514,520,527,533,539,544,551,557,563,569,575,581,587,592,597,604,611,616,621,628,634,640,647,652],{"id":152,"label":451,"issuer":452,"region":161,"url":453,"description":454,"useCases":455,"indexable":284},"EU AI Act","European Union","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":153,"label":457,"issuer":452,"region":161,"url":458,"description":459,"useCases":460,"indexable":284},"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":155,"label":462,"issuer":463,"region":464,"url":465,"description":466,"useCases":467,"indexable":284},"ISO/IEC 42001","ISO and IEC","global","https://www.iso.org/standard/81230.html","The international management system standard for AI.",110,{"id":156,"label":469,"issuer":470,"region":216,"url":471,"description":472,"useCases":473,"indexable":284},"NIST AI Risk Management Framework","NIST","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":475,"label":476,"issuer":452,"region":161,"url":477,"description":478,"useCases":479,"indexable":284},"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":154,"label":481,"issuer":482,"region":161,"url":483,"description":484,"useCases":485,"indexable":284},"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":487,"label":488,"issuer":489,"region":161,"url":490,"description":491,"useCases":492,"indexable":284},"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":494,"label":495,"issuer":496,"region":497,"url":498,"description":499,"useCases":500,"indexable":284},"mas-ai-risk-management","MAS AI risk management guidelines","Monetary Authority of Singapore","asia-pacific","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":502,"label":503,"issuer":504,"region":497,"url":505,"description":506,"useCases":68,"indexable":284},"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.",{"id":508,"label":509,"issuer":510,"region":464,"url":511,"description":512,"useCases":513,"indexable":284},"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":515,"label":516,"issuer":517,"region":216,"url":518,"description":519,"useCases":513,"indexable":284},"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":521,"label":522,"issuer":523,"region":161,"url":524,"description":525,"useCases":526,"indexable":284},"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":528,"label":529,"issuer":530,"region":464,"url":531,"description":532,"useCases":67,"indexable":284},"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.",{"id":534,"label":535,"issuer":452,"region":161,"url":536,"description":537,"useCases":538,"indexable":284},"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":540,"label":541,"issuer":452,"region":161,"url":542,"description":543,"useCases":538,"indexable":284},"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":545,"label":546,"issuer":547,"region":216,"url":548,"description":549,"useCases":550,"indexable":284},"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":552,"label":553,"issuer":452,"region":161,"url":554,"description":555,"useCases":556,"indexable":284},"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":558,"label":559,"issuer":560,"region":216,"url":561,"description":562,"useCases":556,"indexable":284},"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":564,"label":565,"issuer":566,"region":464,"url":567,"description":568,"useCases":556,"indexable":284},"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":570,"label":571,"issuer":452,"region":161,"url":572,"description":573,"useCases":574,"indexable":284},"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":576,"label":577,"issuer":578,"region":216,"url":579,"description":580,"useCases":574,"indexable":284},"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":582,"label":583,"issuer":496,"region":497,"url":584,"description":585,"useCases":586,"indexable":284},"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":588,"label":589,"issuer":452,"region":161,"url":590,"description":591,"useCases":586,"indexable":284},"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":593,"label":594,"issuer":452,"region":161,"url":595,"description":596,"useCases":586,"indexable":284},"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":598,"label":599,"issuer":600,"region":161,"url":601,"description":602,"useCases":603,"indexable":284},"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":605,"label":606,"issuer":607,"region":216,"url":608,"description":609,"useCases":610,"indexable":284},"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":612,"label":613,"issuer":452,"region":161,"url":614,"description":615,"useCases":610,"indexable":284},"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":617,"label":618,"issuer":452,"region":161,"url":619,"description":620,"useCases":437,"indexable":284},"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":622,"label":623,"issuer":624,"region":625,"url":626,"description":627,"useCases":414,"indexable":284},"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":629,"label":630,"issuer":631,"region":161,"url":632,"description":633,"useCases":380,"indexable":284},"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.",{"id":635,"label":636,"issuer":637,"region":161,"url":638,"description":639,"useCases":380,"indexable":284},"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":641,"label":642,"issuer":643,"region":497,"url":644,"description":645,"useCases":646,"indexable":284},"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.",3,{"id":648,"label":649,"issuer":452,"region":161,"url":650,"description":651,"useCases":646,"indexable":284},"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":653,"label":654,"issuer":247,"region":216,"url":655,"description":656,"useCases":646,"indexable":284},"us-fcra","Fair Credit Reporting Act","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.",1790598299895]