[{"data":1,"prerenderedAt":1899},["ShallowReactive",2],{"site-nav-content":3,"hiring-banner-content":177,"blog:\u002Fblog\u002Frbac-for-enterprise-ai":189,"blog-index-copy":651,"blog:\u002Fblog\u002Frbac-for-enterprise-ai:surround":672,"site-cta-content":1880},{"header":4,"productNav":9,"nav":42,"footer":61,"askAI":132,"id":161,"title":162,"archived":163,"authors":164,"badge":164,"body":165,"date":164,"definedTerm":164,"department":164,"description":169,"extension":172,"eyebrow":164,"faqHeader":164,"faqs":164,"footerBand":164,"headline":164,"image":164,"industry":164,"jobType":164,"listed":131,"location":164,"navigation":131,"openRoles":164,"pageLayout":164,"path":173,"relatedHeading":164,"seo":174,"series":164,"sitemap":163,"status":164,"stem":175,"subhead":164,"tags":164,"video":164,"whyJoin":164,"workplaceType":164,"__hash__":176},{"productLabel":5,"loginLabel":6,"contactLabel":7,"contactSalesLabel":8},"Product","Log in","Contact","Get started for free",[10,14,18,22,26,30,34,38],{"label":11,"to":12,"description":13},"Overview","\u002Foverview","Seven layers. One closed loop.",{"label":15,"to":16,"description":17},"Conflux","\u002Fproduct\u002Fconflux","Where your team, workstreams, and agents meet.",{"label":19,"to":20,"description":21},"Agent Teams","\u002Fproduct\u002Fagent-teams","Specialist teams - governed from day one.",{"label":23,"to":24,"description":25},"Lifecycle Graph","\u002Fproduct\u002Flifecycle-graph","Intelligence that compounds across every interaction.",{"label":27,"to":28,"description":29},"Company Wiki","\u002Fproduct\u002Fwiki","Playbooks and policies where expertise stays.",{"label":31,"to":32,"description":33},"Workstreams","\u002Fproduct\u002Fworkstreams","From brief to signed-off deliverable on one canvas.",{"label":35,"to":36,"description":37},"Perception Console","\u002Fproduct\u002Fperception","Ask your whole business in plain English.",{"label":39,"to":40,"description":41},"Governance","\u002Fproduct\u002Fgovernance","Frontier AI you can actually sign off on.",[43,46,49,52,55,58],{"label":44,"to":45},"Models","\u002Fmodels",{"label":47,"to":48},"Pricing","\u002Fpricing",{"label":50,"to":51},"Integrations","\u002Fintegrations",{"label":53,"to":54},"Security","\u002Fsecurity",{"label":56,"to":57},"Partners","\u002Fpartners",{"label":59,"to":60},"Insights","\u002Fblog",{"productHeading":5,"companyHeading":62,"resourcesHeading":63,"legalHeading":64,"docsLabel":65,"docsUrl":66,"statementLines":67,"copyright":70,"companyLinks":71,"resourcesLinks":86,"legalLinks":102,"socialLinks":109,"bottomLinks":119},"Company","Resources","Legal","Docs","https:\u002F\u002Fdocs.gonimbus.ai",[68,69],"Stop training someone else's model.","Control your AI.","© 2026 Nimbus Intelligence, Inc. All rights reserved.",[72,73,74,75,76,78,81,84],{"label":47,"to":48},{"label":50,"to":51},{"label":53,"to":54},{"label":59,"to":60},{"label":77,"to":57},"Partner Program",{"label":79,"to":80},"Careers","\u002Fcareers",{"label":82,"to":83},"System status","\u002Fstatus",{"label":7,"to":85},"\u002Fcontact",[87,90,93,96,99],{"label":88,"to":89},"Glossary","\u002Fglossary",{"label":91,"to":92},"Compare","\u002Fcompare",{"label":94,"to":95},"Evaluate","\u002Fevaluate",{"label":97,"to":98},"Problems","\u002Fproblems",{"label":100,"to":101},"Use cases","\u002Fuse-cases",[103,106],{"label":104,"to":105},"Terms of Service","\u002Fterms",{"label":107,"to":108},"Privacy Policy","\u002Fprivacy",[110,113,116],{"label":111,"href":112},"LinkedIn","https:\u002F\u002Fwww.linkedin.com\u002Fcompany\u002Fgonimbusai\u002F",{"label":114,"href":115},"X","https:\u002F\u002Fx.com\u002Fgonimbusai",{"label":117,"href":118},"Instagram","https:\u002F\u002Fwww.instagram.com\u002Fgonimbus_ai\u002F",[120,122,124,127,128],{"label":121,"to":105},"Terms",{"label":123,"to":108},"Privacy",{"label":125,"to":126},"Compliance","\u002Fcompliance",{"label":82,"to":83},{"label":129,"to":130,"external":131},"LLMs.txt","\u002Fllms.txt",true,{"text":133,"prompt":134,"platforms":135},"Ask AI about Nimbus","I'm researching enterprise intelligence platforms and want to know how Nimbus combines perception, collaboration, and autonomous agents to drive strategic decision-making. Summarize the highlights from Nimbus's website: https:\u002F\u002Fgonimbus.ai",[136,141,146,151,156],{"name":137,"label":138,"icon":139,"hrefPrefix":140},"chatgpt","ChatGPT","simple-icons:openai","https:\u002F\u002Fchatgpt.com\u002F?prompt=",{"name":142,"label":143,"icon":144,"hrefPrefix":145},"perplexity","Perplexity","mdi:magnify","https:\u002F\u002Fwww.perplexity.ai\u002Fsearch\u002Fnew?q=",{"name":147,"label":148,"icon":149,"hrefPrefix":150},"grok","Grok","simple-icons:x","https:\u002F\u002Fx.com\u002Fi\u002Fgrok?text=",{"name":152,"label":153,"icon":154,"hrefPrefix":155},"claude","Claude","simple-icons:anthropic","https:\u002F\u002Fclaude.ai\u002Fnew?q=",{"name":157,"label":158,"icon":159,"hrefPrefix":160},"google-ai","Google AI","simple-icons:google","https:\u002F\u002Fwww.google.com\u002Fsearch?udm=50&aep=11&q=","content\u002Fshared\u002Fnav.md","Site navigation",false,null,{"type":166,"value":167,"toc":168},"minimark",[],{"title":169,"searchDepth":170,"depth":170,"links":171},"",2,[],"md","\u002Fshared\u002Fnav",{"title":162,"description":169},"shared\u002Fnav","Q7sDe7TuGEwhwUMamyeOxjvGlsUaF3Iay9VTW0KaE_U",{"enabled":163,"message":178,"linkLabel":79,"linkHref":80,"id":179,"title":180,"archived":163,"authors":164,"badge":164,"body":181,"date":164,"definedTerm":164,"department":164,"description":169,"extension":172,"eyebrow":164,"faqHeader":164,"faqs":164,"footerBand":164,"headline":164,"image":164,"industry":164,"jobType":164,"listed":131,"location":164,"navigation":131,"openRoles":164,"pageLayout":164,"path":185,"relatedHeading":164,"seo":186,"series":164,"sitemap":163,"status":164,"stem":187,"subhead":164,"tags":164,"video":164,"whyJoin":164,"workplaceType":164,"__hash__":188},"We're hiring! Join the team building the Sentient Enterprise.","content\u002Fshared\u002Fhiring.md","Hiring banner",{"type":166,"value":182,"toc":183},[],{"title":169,"searchDepth":170,"depth":170,"links":184},[],"\u002Fshared\u002Fhiring",{"title":180,"description":169},"shared\u002Fhiring","1zs3boivKda1e-b-hAyuNcmZSKjZUAXmecnwHVgcHzk",{"id":190,"title":191,"archived":163,"authors":192,"badge":197,"body":199,"date":623,"definedTerm":211,"department":164,"description":624,"extension":172,"eyebrow":164,"faqHeader":625,"faqs":628,"footerBand":164,"headline":164,"image":164,"industry":164,"jobType":164,"listed":163,"location":164,"navigation":131,"openRoles":164,"pageLayout":164,"path":644,"relatedHeading":164,"seo":645,"series":646,"sitemap":131,"status":164,"stem":647,"subhead":164,"tags":648,"video":164,"whyJoin":164,"workplaceType":164,"__hash__":650},"content\u002Fblog\u002Frbac-for-enterprise-ai.md","What is RBAC for enterprise AI, and why should you care?",[193],{"name":194,"role":195,"to":196},"Zac Radbone","Co-Founder and CCO","https:\u002F\u002Fgonimbus.ai",{"label":198},"Explainer",{"type":166,"value":200,"toc":613},[201,205,213,221,226,234,237,250,253,266,269,276,280,283,291,304,307,313,324,330,397,400,408,412,415,418,448,451,455,462,465,478,482,485,488,494,500,503,507,518,521,534,538,563,567],[202,203,204],"p",{},"The campaign ended in March. The agency guest is still in the room in June. Nobody meant to leave them there. The invite was a courtesy, the workstream was busy, and the login looked harmless because the model was only supposed to summarise. Then a member pasted a CRM export into the thread so the assistant could “just take a look.” The guest never needed a production password. They inherited a session that could already see customer rows. When security asks who could see the file, the honest answer is a display name and a date that nobody reviewed.",[202,206,207,208,212],{},"That is the enterprise AI access problem in one room. Role-based access control — ",[209,210,211],"strong",{},"RBAC"," — is who is allowed to do what. For people it is familiar: guests, members, admins; remove the contractor when the project ends. For a model, the “who” is the session acting with someone’s credentials — reading files those credentials can reach, and sometimes proposing a change to a live system. A fluent answer can still be the wrong change in the wrong place.",[202,214,215,220],{},[216,217,219],"a",{"href":218},"\u002Fblog\u002Fwhat-is-ai-governance\u002F","What is AI governance"," is the parent definition. RBAC is the people-and-tools half. This essay is not a claim that one product has solved it. It is why the noun exists, what fails when a chatbot login is treated as the whole answer, and a way to start this quarter.",[222,223,225],"h2",{"id":224},"why-the-noun-exists","Why the noun exists",[202,227,228,229,233],{},"Classic RBAC, described by Ferraiolo and Kuhn in a ",[216,230,232],{"href":231},"https:\u002F\u002Fcsrc.nist.gov\u002Ffiles\u002Fpubs\u002Fconference\u002F1992\u002F10\u002F13\u002Frolebased-access-controls\u002Ffinal\u002Fdocs\u002Fferraiolo-kuhn-92.pdf","NIST paper"," in 1992, assigns permissions to roles, then roles to people. The point was never elegance. It was that individual user-to-permission maps rot the moment headcount moves. A role is a job you can revoke.",[202,235,236],{},"Enterprise AI adds three questions the 1992 paper did not have to ask:",[238,239,240,244,247],"ul",{},[241,242,243],"li",{},"Which jobs and files can this person — and this model, in this session — see?",[241,245,246],{},"Which tools can it call?",[241,248,249],{},"If it can change a live system, who must approve, and is that approval stored?",[202,251,252],{},"A chatbot login answers a fourth, narrower question: may this person talk to the bot? That is necessary. It is not the same as answering the three above. The login is a door. RBAC is the floor plan.",[202,254,255,256,260,261,265],{},"NIST’s ",[216,257,259],{"href":258},"https:\u002F\u002Fwww.nist.gov\u002Fitl\u002Fai-risk-management-framework","AI Risk Management Framework"," puts the same idea in public language: govern, map, measure, manage. You cannot measure a write you cannot attribute. ",[216,262,264],{"href":263},"https:\u002F\u002Fcsrc.nist.gov\u002Fpubs\u002Fsp\u002F800-207\u002Ffinal","SP 800-207"," on zero trust is the companion: do not assume a session is trusted because it signed in. A model that inherited a sales manager’s token is still a session, not a person you can walk to a desk.",[202,267,268],{},"Guests, members, and admins are who is on the job. The model side is which tools that session may call. Both belong in RBAC.",[202,270,271,275],{},[216,272,274],{"href":273},"https:\u002F\u002Fwww.mckinsey.com\u002Fcapabilities\u002Fquantumblack\u002Four-insights\u002Fthe-state-of-ai","McKinsey’s State of AI"," keeps finding the same organisational pattern: use is ordinary, and operating redesign is not. Eighty-eight percent of respondents say their organisations use AI in at least one function; nearly two-thirds remain in experimentation or piloting. Programmes that scale a model into every function without scaling roles onto every job will discover the guest-and-paste problem as an incident, not as a design review.",[222,277,279],{"id":278},"the-hidden-failure","The hidden failure",[202,281,282],{},"The failure is not usually a Hollywood breach. It is a customer record changed without a name next to the change, or a contractor who still sees a workstream after the project ended.",[202,284,285,286,290],{},"IBM’s ",[216,287,289],{"href":288},"https:\u002F\u002Fnewsroom.ibm.com\u002F2025-07-30-ibm-report-13-of-organizations-reported-breaches-of-ai-models-or-applications,-97-of-which-reported-lacking-proper-ai-access-controls","Cost of a Data Breach research"," found that one in five organisations studied reported a breach involving shadow AI, and that high levels of shadow AI were associated with about $670,000 in higher average breach costs. Only 37 percent had policies to manage AI or detect unsanctioned use. You do not need those numbers to justify RBAC. You need one guest who outlived the job, or one unofficial tool that never received a role at all.",[202,292,293,294,298,299,303],{},"Microsoft and LinkedIn’s ",[216,295,297],{"href":296},"https:\u002F\u002Fwww.microsoft.com\u002Fen-us\u002Fworklab\u002Fwork-trend-index\u002Fai-at-work-is-here-now-comes-the-hard-part","Work Trend Index"," found that 78 percent of AI users bring their own tools. That is ",[216,300,302],{"href":301},"\u002Fblog\u002Fwhat-is-shadow-ai\u002F","shadow AI",": useful, and outside the roles you think you assigned. The official RBAC matrix on the sanctioned chatbot does not cover the personal account on the phone. A board that has never seen the unofficial path is not conservative. It is uninformed.",[202,305,306],{},"Three failures follow when the login is treated as the control.",[202,308,309,312],{},[209,310,311],{},"Inherited sight."," A member pastes a file the guest should never have seen. The model summarises it for everyone in the room. Revoking the guest login later does not un-see the export. Roles that follow the job — not only the person — are how you close that gap. Roles that follow only the seat leave the file in the history.",[202,314,315,318,319,323],{},[209,316,317],{},"Unsigned writes."," Write-back means the AI changes a live system. Fail-closed means if nobody approves, nothing happens. A prompt that says “ask first” is not fail-closed. If the model can post because the connector was on for retrieval, you have given a session the rights of a clerk without giving it the duties of a clerk. ",[216,320,322],{"href":321},"\u002Fblog\u002Fwhat-is-write-back-governance\u002F","Write-back governance"," is the checklist for that path.",[202,325,326,329],{},[209,327,328],{},"Orphaned access."," The campaign ends. The guest remains. The shared password remains. The “temporary” API key remains. Offboarding that disables a human account and leaves the model’s tool grants untouched is half an offboarding.",[331,332,333,349],"table",{},[334,335,336],"thead",{},[337,338,339,343,346],"tr",{},[340,341,342],"th",{},"What you granted",[340,344,345],{},"What you think you granted",[340,347,348],{},"What an auditor will ask",[350,351,352,364,375,386],"tbody",{},[337,353,354,358,361],{},[355,356,357],"td",{},"A chatbot login",[355,359,360],{},"Access control",[355,362,363],{},"Who could see this job last quarter?",[337,365,366,369,372],{},[355,367,368],{},"Read on a connector",[355,370,371],{},"Safety",[355,373,374],{},"Which customer files entered whose session?",[337,376,377,380,383],{},[355,378,379],{},"A shared agency seat",[355,381,382],{},"Convenience",[355,384,385],{},"When was that seat removed from this roster?",[337,387,388,391,394],{},[355,389,390],{},"A prompt that says ask first",[355,392,393],{},"Oversight",[355,395,396],{},"Show the stored no, or show that write is off",[202,398,399],{},"If AI only summarises public wiki pages, the stakes are lower. You can still use roles so the wiki is not everyone’s dump of customer data. You should care if you have guests on a job, if AI can write to CRM or finance systems, or if someone might later ask who approved a machine-initiated change.",[202,401,402,403,407],{},"Where ",[216,404,406],{"href":405},"https:\u002F\u002Fartificialintelligenceact.eu\u002Farticle\u002F14\u002F","Article 14 of the EU AI Act"," applies, human oversight is not a footer. People must be able to interpret the output and interrupt or refuse the system. A role that can open the chat but cannot refuse the write is not that oversight. You do not have to be in scope of the Act for the operational question to be useful. Customers and boards will ask it in other words.",[222,409,411],{"id":410},"how-to-apply-it-when-ai-can-change-records","How to apply it when AI can change records",[202,413,414],{},"Write-back means the AI changes a live system. A payload is the exact change, shown before it goes out. A signer is a person whose role owns that class of change — not “the channel,” not “the bot user.”",[202,416,417],{},"A practical sequence:",[419,420,421,427,430,438,441],"ol",{},[241,422,423,424,426],{},"Keep the model from writing until you can name the object class and the signer. ",[216,425,322],{"href":321}," is the checklist. Read-only by default is the honest starting state, not a temporary embarrassment.",[241,428,429],{},"For each write class, name the approver role. Finance on a journal. The account owner on a discount. Legal on a clause that will be sent. If you cannot name the role, the class is not ready.",[241,431,432,433,437],{},"Store the payload and the decision so you can reopen them. ",[216,434,436],{"href":435},"\u002Fblog\u002Fwhat-auditors-are-asking-for\u002F","What auditors are asking for"," is the evidence pack: a named person, a frozen change, a stored yes or no.",[241,439,440],{},"When someone leaves the job, remove them from the roster the same week. Review guest access on a cadence, not when security asks.",[241,442,443,444,447],{},"List the unofficial tools. ",[216,445,446],{"href":301},"Shadow AI"," is what happens when the unofficial path never got those roles. A ban without a usable sanctioned path moves the work to a phone.",[202,449,450],{},"Decision History — a log of actor, tier, and timestamp — can show who signed a proposal. It is not the statutory archive, the ledger, or the customer file the company already keeps. Do not buy a trail and retire the system of record. Counsel names where the books live. The product trail is how you walk a single change without reconstructing a chat.",[222,452,454],{"id":453},"what-to-ask-a-vendor","What to ask a vendor",[202,456,457,458,461],{},"A short list of demo questions lives in ",[216,459,460],{"href":435},"what auditors are asking for",". In one sentence: can they show who could see a job, which tool ran, and who approved a write — without a screenshot hunt?",[202,463,464],{},"Ask the concrete version. Pick a guest; show the files that session can reach; remove them and show the next run cannot. Name one connector; show it read-only; attempt a write and show the failure. Propose a change; refuse it; show the stored no. Ask whether Sales can see a payroll connector. “The workspace” is not an answer.",[202,466,467,468,472,473,477],{},"The ",[216,469,471],{"href":470},"https:\u002F\u002Feur-lex.europa.eu\u002Flegal-content\u002FEN\u002FTXT\u002F?uri=CELEX:32024R1689","EU AI Act"," and ",[216,474,476],{"href":475},"https:\u002F\u002Fwww.iso.org\u002Fstandard\u002F81230.html","ISO\u002FIEC 42001"," are why those questions show up in procurement. You do not have to implement every clause on day one. You do need an answer you could give an auditor. ISO 42001 is a management system. It does not replace a named person on a payload. You can be on a certification path and still have an admin token on a model.",[222,479,481],{"id":480},"how-to-start-this-quarter","How to start this quarter",[202,483,484],{},"You need one job, one roster, and one write class that stays off until the signer is named.",[202,486,487],{},"This month, pick a cross-team job that already has guests. Write down who is on it and why. Remove anyone whose reason is “they were copied in.” Attach the two files the job needs. Keep connectors read-only.",[202,489,490,491,493],{},"Next month, name the first write class you might enable and the role that may sign it. Run one refusal in a sandbox. Walk ",[216,492,460],{"href":435}," against that change. Fill blanks as findings.",[202,495,496,497,499],{},"Then list unofficial tools that touch customer files, unpublished figures, or source code. You will not eliminate ",[216,498,302],{"href":301}," with a memo. You will learn which sanctioned path is too slow — people leave the role system you designed because it cannot finish the job.",[202,501,502],{},"Score any product as an access review: named roster, named tools, write off until a signer exists, a stored no you can reopen. If the demo starts by sharing a password into a consumer model, you are scoring a chat seat.",[222,504,506],{"id":505},"how-this-shows-up-in-nimbus","How this shows up in Nimbus",[202,508,509,510,513,514,517],{},"Nimbus is a Collaborative AI operating system: humans and agent teams on the same workstream, with roles on the job rather than only on a chat login. Connectors default to read-only. Writes wait for a named approval. The platform reaches 2,000+ business tools; a connection is not permission to mutate the source. ",[216,511,39],{"href":512},"\u002Fproduct\u002Fgovernance\u002F"," is the product surface. ",[216,515,53],{"href":516},"\u002Fsecurity\u002F"," states the compliance lines as they stand: GDPR today; ISO 27001 and SOC 2 Type 1 independent audits underway — not certificates for a slide.",[202,519,520],{},"Decision History logs actor, tier, and timestamp. It is a trail you can filter, not the statutory books-and-records store. Hugh Woods Ballard at Day Webster wanted specialists and people in the same place, and a way to see who approved what. Andy at Digby Fine English wanted humans still in the loop before anything commits. That is RBAC as a daily habit, not a policy PDF.",[202,522,523,524,528,529,533],{},"For how teams share the job once access is clear, see ",[216,525,527],{"href":526},"\u002Fblog\u002Fwhat-is-collaborative-ai\u002F","what is collaborative AI",". For where the decision should live after the thread ends, see ",[216,530,532],{"href":531},"\u002Fblog\u002Fsearch-is-not-memory\u002F","search is not memory",".",[222,535,537],{"id":536},"related-reading","Related reading",[238,539,540,544,549,553,558],{},[241,541,542],{},[216,543,219],{"href":218},[241,545,546],{},[216,547,548],{"href":321},"What is write-back governance",[241,550,551],{},[216,552,436],{"href":435},[241,554,555],{},[216,556,557],{"href":301},"What is shadow AI",[241,559,560],{},[216,561,562],{"href":526},"What is collaborative AI",[222,564,566],{"id":565},"references","References",[238,568,569,574,579,584,589,594,599,604,609],{},[241,570,571],{},[216,572,573],{"href":231},"Ferraiolo and Kuhn, Role-Based Access Controls, NIST, 1992",[241,575,576],{},[216,577,578],{"href":258},"NIST AI Risk Management Framework",[241,580,581],{},[216,582,583],{"href":263},"NIST SP 800-207, Zero Trust Architecture",[241,585,586],{},[216,587,588],{"href":273},"McKinsey, The State of AI",[241,590,591],{},[216,592,593],{"href":288},"IBM, Cost of a Data Breach research release, 30 July 2025",[241,595,596],{},[216,597,598],{"href":296},"Microsoft and LinkedIn, 2024 Work Trend Index",[241,600,601],{},[216,602,603],{"href":405},"EU AI Act, Article 14, human oversight",[241,605,606],{},[216,607,608],{"href":470},"EU AI Act, Regulation (EU) 2024\u002F1689",[241,610,611],{},[216,612,476],{"href":475},{"title":169,"searchDepth":170,"depth":170,"links":614},[615,616,617,618,619,620,621,622],{"id":224,"depth":170,"text":225},{"id":278,"depth":170,"text":279},{"id":410,"depth":170,"text":411},{"id":453,"depth":170,"text":454},{"id":480,"depth":170,"text":481},{"id":505,"depth":170,"text":506},{"id":536,"depth":170,"text":537},{"id":565,"depth":170,"text":566},"2026-08-27","A shared chatbot login is not access control when a model can see customer files — or write to a live system — under someone else’s credentials.",{"eyebrow":626,"title":627},"Short answers","Roles when the user is a model",[629,632,635,638,641],{"question":630,"answer":631},"Is a shared chatbot login the same as RBAC?","No. A shared login says who can open the chat. RBAC says who can see which jobs, which tools, and which live systems — and whether the model may write at all.",{"question":633,"answer":634},"Do we need RBAC if AI is read-only?","You still need it for what the model can see. Read-only reduces the chance of a bad write. It does not decide which customer files belong in whose session.",{"question":636,"answer":637},"Where should we start?","Name who can approve a change to a live system, keep AI from writing until that is clear, and list unofficial tools. The auditors guide on this site is a first evidence pack.",{"question":639,"answer":640},"Does a Decision History log replace our statutory archive?","No. A log of actor, tier, and timestamp shows who signed a proposal. It is not the books-and-records store, the ledger, or the customer archive the company already keeps. Counsel still names the system of record.",{"question":642,"answer":643},"What does Article 14 of the EU AI Act change about roles?","Where it applies, human oversight must be able to interpret an output and interrupt or refuse the system. A chatbot login that cannot refuse a write is not that oversight, whatever the role name on the seat.","\u002Fblog\u002Frbac-for-enterprise-ai",{"title":191,"description":624},"explainer","blog\u002Frbac-for-enterprise-ai",[646,211,649],"access","pWGw22381C9pmnKzx5CaVeOJT2XWpLs_5UXzzZ7v6E4",{"hero":652,"id":654,"title":655,"archived":163,"authors":164,"badge":164,"body":656,"date":164,"definedTerm":164,"department":164,"description":660,"extension":172,"eyebrow":661,"faqHeader":164,"faqs":164,"footerBand":662,"headline":164,"image":164,"industry":164,"jobType":164,"listed":131,"location":164,"navigation":131,"openRoles":164,"pageLayout":164,"path":60,"relatedHeading":668,"seo":669,"series":164,"sitemap":131,"status":164,"stem":670,"subhead":164,"tags":164,"video":164,"whyJoin":164,"workplaceType":164,"__hash__":671},{"filename":653},"u2221455217_Flat_design_of_a_futuristic_minimalist_landscape__5d589295-cdea-4ea9-a262-be766881accf_1.png","content\u002Fblog\u002Findex.md","Exploring the future of intelligence.",{"type":166,"value":657,"toc":658},[],{"title":169,"searchDepth":170,"depth":170,"links":659},[],"Deep dives into pre-cognitive intelligence, sentient enterprises, and the evolving landscape of AI-driven business transformation.","Latest Research",{"headline":663,"description":664,"primaryLabel":665,"primaryTo":666,"secondaryLabel":667,"secondaryTo":12},"Stay at the frontier.","Subscribe for product updates and new insights.","Subscribe","\u002Fnewsletter","Explore the platform","More research",{"title":655,"description":660},"blog\u002Findex","BFSWGYO9bcTlaulivKYWyg08_DJHsdGg3OC6g_CG1Hw",[673,1170],{"id":674,"title":675,"archived":163,"authors":676,"badge":679,"body":680,"date":1148,"definedTerm":164,"department":164,"description":1149,"extension":172,"eyebrow":164,"faqHeader":1150,"faqs":1152,"footerBand":164,"headline":164,"image":164,"industry":164,"jobType":164,"listed":163,"location":164,"navigation":131,"openRoles":164,"pageLayout":164,"path":1162,"relatedHeading":164,"seo":1163,"series":646,"sitemap":131,"status":164,"stem":1164,"subhead":164,"tags":1165,"video":164,"whyJoin":164,"workplaceType":164,"__hash__":1169},"content\u002Fblog\u002Fwhat-is-a-company-wiki-for-ai-agents.md","What is a Company Wiki for AI Agents",[677],{"name":678,"to":196},"Nimbus Research",{"label":198},{"type":166,"value":681,"toc":1122},[682,689,692,700,703,707,710,744,747,796,799,803,809,812,844,849,864,874,880,886,894,898,905,912,919,923,926,933,936,950,952,957,964,967,982,986,990,997,1001,1004,1008,1011,1015,1018,1022,1032,1036,1039,1043,1046,1050,1053,1057,1064,1068,1074,1078,1085,1089,1092,1094,1104,1108],[202,683,684,685,688],{},"A company wiki for AI agents is the ",[209,686,687],{},"official playbook the AI must follow",": owned, versioned, and scoped — not a pile of old Drive files that search might find.",[202,690,691],{},"Human wikis (Confluence, Notion, SharePoint) were built for people: pages, comments, “someone should update this.” Agent wikis have a harder job. Models will obey the loudest chunk in the prompt unless you separate kinds of text on purpose.",[202,693,694,695,699],{},"If the discount floor lives in a slide, a Slack rumour, and last year’s deck, an assistant asked to draft an exception will pick whichever document ",[696,697,698],"em",{},"sounds"," closest to the question. That is not policy. That is folklore with a search box.",[202,701,702],{},"The distinction is easy to miss because both surfaces look like “knowledge.” One is a library. The other is a constitution. An agent that can retrieve every file still does not know which file is currently in force unless the runtime loads asserted policy on purpose.",[222,704,706],{"id":705},"words-youll-hear","Words you’ll hear",[202,708,709],{},"Keep three kinds of text apart:",[238,711,712,718,733],{},[241,713,714,717],{},[209,715,716],{},"Asserted."," What the company currently wants. Owned. Dated. Scoped. This is the wiki. At work, this is the pricing floor, the refund rule, the journal-posting checklist, the approved customer language. If legal updated it on Tuesday, the agent must cite Tuesday’s version on Wednesday — not the semantically similar PDF from 2023.",[241,719,720,723,724,728,729,732],{},[209,721,722],{},"Retrieved."," What exists in systems. Possibly stale or contradictory. That is ",[216,725,727],{"href":726},"\u002Fblog\u002Fwhat-is-enterprise-rag\u002F","enterprise RAG",": look up authorised files, then answer. Lookup is not the same as “this is policy.” At work, retrieval is last quarter’s board pack, a ticket thread, a contract PDF. Those documents may be true as ",[696,730,731],{},"records",". They are not automatically the rule you want the agent to follow next.",[241,734,735,738,739,743],{},[209,736,737],{},"Decided."," What we already approved in a run, stored on the ",[216,740,742],{"href":741},"\u002Fblog\u002Fwhat-is-a-lifecycle-graph\u002F","lifecycle graph",". A signed exception should not silently overwrite the playbook for everyone else. At work, this is “this renewal was allowed 18% because of a named exception.” That fact belongs on the decision chain. It does not become the new global discount floor unless a human promotes it into the wiki.",[202,745,746],{},"Other terms you will hear in vendor decks and internal Slack, and how they actually show up:",[238,748,749,755,766,772,778,784,790],{},[241,750,751,754],{},[209,752,753],{},"Vault."," A scoped partition of knowledge (finance vs people ops) with role-based access. At work, finance’s close checklist should not ride along in a recruiting workstream “just in case the model finds it useful.”",[241,756,757,760,761,765],{},[209,758,759],{},"Citation."," The answer names the page and version — ",[762,763,764],"code",{},"pricing v4.2"," — not “the wiki.” At work, an auditor or a new manager should be able to open the same page the agent used, not reconstruct a vibe.",[241,767,768,771],{},[209,769,770],{},"Conflict rule."," If Drive contradicts the wiki, the wiki wins unless a human promotes a change. At work, this is the only way a retrieval-heavy assistant stops treating the loudest PDF as law.",[241,773,774,777],{},[209,775,776],{},"Authority marker."," Labels such as policy, draft, archive, and local exception. Drafts must not load as binding context.",[241,779,780,783],{},[209,781,782],{},"Owner."," A named role, not “the AI team.” The discount floor is owned by revenue operations or finance, not by whoever last edited a Notion page.",[241,785,786,789],{},[209,787,788],{},"Review cadence."," A date when the page is re-checked. Silence becomes folklore.",[241,791,792,795],{},[209,793,794],{},"Scope."," Which jobs may load this page. People-ops rules are not in the go-to-market context by default.",[202,797,798],{},"Most “knowledge bases,” custom GPTs, and giant system prompts fail here because they are either too global (one constitution for every department) or too private (each user pastes rules into a personal assistant). Neither is owned. Neither is maintained.",[222,800,802],{"id":801},"why-you-should-care","Why you should care",[202,804,805,806,808],{},"When official policy is unusable, employees ask consumer models to invent policy. See ",[216,807,557],{"href":301},". The unofficial tool will synthesise a refund rule or a customer commitment from whatever was pasted. The company still owns the result.",[202,810,811],{},"It affects you if:",[238,813,814,820,830,836],{},[241,815,816,819],{},[209,817,818],{},"Numbers in playbooks disagree"," with numbers in CRM, and nobody can say which is official.",[241,821,822,825,826,533],{},[209,823,824],{},"People leave."," Tacit knowledge — the hallway version of the rule — leaves with them. See ",[216,827,829],{"href":828},"\u002Fblog\u002Fwhat-is-institutional-memory-in-enterprise-ai\u002F","What is institutional memory in enterprise AI",[241,831,832,835],{},[209,833,834],{},"Legal or finance must cite a version",", not a vibe.",[241,837,838,841,842,533],{},[209,839,840],{},"Agents can propose writes."," A model that can change CRM without a binding playbook is improvising in production. See ",[216,843,548],{"href":321},[845,846,848],"h3",{"id":847},"what-changes-by-role","What changes by role",[202,850,851,854,855,858,859,863],{},[209,852,853],{},"Finance."," The wiki is where recognition rules, posting checklists, and materiality thresholds live as asserted text. Retrieval of last year’s close pack is not a substitute. If an agent drafts a journal from a Slack thread that contradicts the close checklist, finance needs the conflict rule to fire ",[696,856,857],{},"before"," a named signer is asked to approve. Token spend also changes: re-deriving the same policy from a pile of PDFs every run is how ",[216,860,862],{"href":861},"\u002Fblog\u002Fwhat-is-ai-token-economics\u002F","token economics"," inflate without improving the artefact.",[202,865,866,869,870,873],{},[209,867,868],{},"Legal."," Approved language, retention classes, and “do not say” lists belong in asserted pages with owners. A retrieved contract is evidence of what was signed with ",[696,871,872],{},"that"," counterparty. It is not the company’s current standard terms. Legal also cares that citations name a version. “According to our documents” is not a defence if those documents include drafts.",[202,875,876,879],{},[209,877,878],{},"Operations."," Runbooks, escalation thresholds, and supplier exception rules need to be loadable as the current procedure, nothe closest matching incident write-up. Ops already knows that a stale SOP is worse than no SOP, because people follow it. Agents do the same, faster.",[202,881,882,885],{},[209,883,884],{},"Go-to-market."," Discount floors, win\u002Floss taxonomies, and approved competitive language are the pages that stop an assistant inventing a concession. GTM also feels the scope problem first: a “help me close this” chat that can see every playbook in the company will mix people-ops rules, finance forecasts, and last year’s campaign into one fluent paragraph.",[202,887,888,891,892,533],{},[209,889,890],{},"Security."," Vaults and least-privilege loading are access control. Indexing every SharePoint site into a single “brain” is a new store of sensitive data. Security’s question is not “does the model know enough?” It is “which pages is this job allowed to load, and can we prove it?” GDPR-style purpose limitation still applies when the reader is an AI. See ",[216,893,219],{"href":218},[845,895,897],{"id":896},"what-people-get-wrong","What people get wrong",[202,899,900,901,904],{},"The common failure is treating ",[209,902,903],{},"search as policy",". Teams export Confluence into a vector index, label it “the brain,” and congratulate themselves for grounding. Search will surface the outdated note because it is semantically close to the question. Grounding on the wrong document is still grounding. It is just grounding on folklore.",[202,906,907,908,911],{},"The second failure is the ",[209,909,910],{},"personal constitution",": each power user pastes rules into a custom GPT. Those rules are not org-owned, not scoped per job, and not cited as a version in an audit. When two users paste different discount floors, the company has two unofficial policies.",[202,913,914,915,918],{},"The third failure is the ",[209,916,917],{},"mega-prompt",". One global instruction block tries to encode every department. It is never current. It cannot be scoped. It cannot be reviewed by the owner of a single domain. It also burns tokens on every call.",[845,920,922],{"id":921},"what-good-looks-like-versus-what-fails","What good looks like versus what fails",[202,924,925],{},"A good agent wiki has authority markers (policy vs draft vs archive), scope (people-ops rules are not in the go-to-market context by default), versions, named owners, a review cadence, and tables for numbers. Numbers belong in tables because prose rounds them. Agents will quote the table if you give them one.",[202,927,928,929,932],{},"A good wiki is also ",[209,930,931],{},"written for two audiences",": the human who must own the page, and the agent that must cite it. Humans need headings and owners. Agents need unambiguous numbers and conflict rules.",[202,934,935],{},"A bad agent wiki is an export of Confluence into a search index. The intranet wiki remains useful for humans. It is still not binding on agents unless the runtime loads a controlled subset. Connection is not the same as “everything in Confluence is policy.”",[202,937,938,939,941,942,944,945,949],{},"Adjacent ideas: retrieval without assertion is ",[216,940,727],{"href":726},". Decisions without a playbook are a ",[216,943,742],{"href":741}," with nothing to cite. A job that loads the wrong vault is a ",[216,946,948],{"href":947},"\u002Fblog\u002Fwhat-is-an-ai-workstream\u002F","workstream"," with the wrong attachments.",[222,951,506],{"id":505},[202,953,467,954,956],{},[209,955,27],{}," is the asserted policy layer every agent team must treat as binding. Workstreams subscribe to wiki sections so scope is enforced at runtime. Pages connect to runs and approvals on the Lifecycle Graph.",[202,958,959,960,963],{},"The wiki is not a second search engine. Connectors remain the path to live systems, and they default to read-only. Retrieval of Drive or CRM is still retrieval. The wiki is what those reads are interpreted ",[696,961,962],{},"against",". When a write is proposed, the named signer should see the playbook version the draft claims to follow.",[202,965,966],{},"Perception can ask what the current playbook says, and which run last cited it.",[202,968,969,970,974,975,978,979,533],{},"See ",[216,971,973],{"href":972},"\u002Fproduct\u002Fwiki\u002F","Wiki",". For the job that loads a subset of pages, see ",[216,976,31],{"href":977},"\u002Fproduct\u002Fworkstreams\u002F",". For the chain that records which version was used, see ",[216,980,23],{"href":981},"\u002Fproduct\u002Flifecycle-graph\u002F",[222,983,985],{"id":984},"questions-people-actually-ask","Questions people actually ask",[845,987,989],{"id":988},"isnt-this-just-confluence","Isn’t this just Confluence?",[202,991,992,993,996],{},"Confluence is a human wiki. An agent wiki is a ",[696,994,995],{},"binding"," subset: owned, versioned, scoped, and loaded on purpose. You can connect Confluence into that layer. Connection is not the same as “everything in Confluence is policy.”",[845,998,1000],{"id":999},"cant-we-just-search-drive","Can’t we just search Drive?",[202,1002,1003],{},"Search is retrieval. Retrieval finds what exists. It does not decide what the company currently wants. If Drive contains three discount floors, search will return the closest one, not the official one.",[845,1005,1007],{"id":1006},"what-if-the-wiki-is-wrong","What if the wiki is wrong?",[202,1009,1010],{},"Then a human updates it, with a version and an owner. Do not let a one-off exception silently become the new global rule. Promote the change; do not hope the next retrieval will “learn.”",[845,1012,1014],{"id":1013},"how-is-this-different-from-a-custom-gpts-instructions","How is this different from a custom GPT’s instructions?",[202,1016,1017],{},"Instructions in a personal GPT are not org-owned, not scoped per job, and not cited as a version in an audit. Two users can ship two unofficial policies without anyone noticing until a customer is told the wrong thing.",[845,1019,1021],{"id":1020},"do-we-need-a-wiki-if-we-already-have-rag","Do we need a wiki if we already have RAG?",[202,1023,1024,1025,1028,1029,533],{},"Yes, if agents will act. RAG reduces invention on ",[696,1026,1027],{},"existing"," files. It does not mark which file is in force. Without assertion, retrieval-augmented generation is retrieval-augmented folklore. See ",[216,1030,1031],{"href":726},"What is enterprise RAG",[845,1033,1035],{"id":1034},"who-should-own-wiki-pages","Who should own wiki pages?",[202,1037,1038],{},"The same function that owns the analogue rule. Pricing belongs to revenue operations or finance. Employment language belongs to people ops and legal. “The AI team” is a coordinator, not a policy owner.",[845,1040,1042],{"id":1041},"how-often-should-pages-be-reviewed","How often should pages be reviewed?",[202,1044,1045],{},"On a cadence that matches how often the rule changes, plus a hard date so silence is visible. A discount floor that never expires is how last year’s promotion becomes this year’s default.",[845,1047,1049],{"id":1048},"what-should-we-put-in-tables-versus-prose","What should we put in tables versus prose?",[202,1051,1052],{},"Numbers, thresholds, codes, and “never \u002F always” lists belong in tables. Narrative belongs in prose. Agents quote tables more reliably than they extract a number buried in a paragraph.",[845,1054,1056],{"id":1055},"can-one-wiki-serve-the-whole-company","Can one wiki serve the whole company?",[202,1058,1059,1060,1063],{},"One ",[696,1061,1062],{},"product",", many vaults. A single unscoped corpus recreates the god workspace. Finance close pages and recruiting pages should not share a default context.",[845,1065,1067],{"id":1066},"how-do-exceptions-work-without-rewriting-the-playbook","How do exceptions work without rewriting the playbook?",[202,1069,1070,1071,533],{},"Record the exception on the decision chain — who signed, which page version, which record — and leave the playbook intact unless a human promotes a change. See ",[216,1072,1073],{"href":741},"What is a lifecycle graph",[845,1075,1077],{"id":1076},"will-a-better-model-make-the-wiki-unnecessary","Will a better model make the wiki unnecessary?",[202,1079,1080,1081,533],{},"No. Stronger models are better at sounding like policy. That makes an unowned corpus more dangerous, not less. Model routing can send interpretation to a stronger model; it cannot invent an owner. See ",[216,1082,1084],{"href":1083},"\u002Fblog\u002Fwhat-is-model-routing\u002F","What is model routing",[845,1086,1088],{"id":1087},"how-does-this-relate-to-access-control","How does this relate to access control?",[202,1090,1091],{},"Loading a page is still processing. A recruiting workstream should not load compensation policy “because it might help.” Vaults and workstream subscriptions keep that promise in software rather than in a PDF.",[222,1093,537],{"id":536},[202,1095,1096,1098,1099,1101,1102,533],{},[216,1097,1031],{"href":726},", ",[216,1100,829],{"href":828},", and ",[216,1103,219],{"href":218},[222,1105,1107],{"id":1106},"sources","Sources",[238,1109,1110,1116],{},[241,1111,1112],{},[216,1113,1115],{"href":1114},"https:\u002F\u002Fico.org.uk\u002Ffor-organisations\u002Fuk-gdpr-guidance-and-resources\u002Fartificial-intelligence\u002F","ICO, AI and data protection",[241,1117,1118],{},[216,1119,1121],{"href":1120},"https:\u002F\u002Feur-lex.europa.eu\u002Feli\u002Freg\u002F2016\u002F679\u002Foj","GDPR",{"title":169,"searchDepth":170,"depth":170,"links":1123},[1124,1125,1131,1132,1146,1147],{"id":705,"depth":170,"text":706},{"id":801,"depth":170,"text":802,"children":1126},[1127,1129,1130],{"id":847,"depth":1128,"text":848},3,{"id":896,"depth":1128,"text":897},{"id":921,"depth":1128,"text":922},{"id":505,"depth":170,"text":506},{"id":984,"depth":170,"text":985,"children":1133},[1134,1135,1136,1137,1138,1139,1140,1141,1142,1143,1144,1145],{"id":988,"depth":1128,"text":989},{"id":999,"depth":1128,"text":1000},{"id":1006,"depth":1128,"text":1007},{"id":1013,"depth":1128,"text":1014},{"id":1020,"depth":1128,"text":1021},{"id":1034,"depth":1128,"text":1035},{"id":1041,"depth":1128,"text":1042},{"id":1048,"depth":1128,"text":1049},{"id":1055,"depth":1128,"text":1056},{"id":1066,"depth":1128,"text":1067},{"id":1076,"depth":1128,"text":1077},{"id":1087,"depth":1128,"text":1088},{"id":536,"depth":170,"text":537},{"id":1106,"depth":170,"text":1107},"2026-08-17","A company wiki for AI agents is the official playbook the AI must follow: versioned, owned, and scoped.",{"eyebrow":626,"title":1151},"The playbook the agent must follow",[1153,1156,1159],{"question":1154,"answer":1155},"What is a company wiki for AI agents?","The official playbook the AI must follow: versioned, owned, and scoped.",{"question":1157,"answer":1158},"Is a Drive folder that wiki?","No. A pile of old files the search might find is not an official source. The agent needs a playbook someone owns.",{"question":1160,"answer":1161},"What happens when the playbook and a prompt disagree?","The playbook wins. A prompt is a request. The wiki is the rule.","\u002Fblog\u002Fwhat-is-a-company-wiki-for-ai-agents",{"title":675,"description":1149},"blog\u002Fwhat-is-a-company-wiki-for-ai-agents",[646,1166,1167,1168],"wiki","agents","knowledge","ukMrOkIBq5q4AJTTJsUqSZDnFr8zgd5-UX2rdDlsr24",{"id":1171,"title":1172,"archived":163,"authors":1173,"badge":1175,"body":1176,"date":1855,"definedTerm":164,"department":164,"description":1856,"extension":172,"eyebrow":164,"faqHeader":1857,"faqs":1860,"footerBand":164,"headline":164,"image":164,"industry":164,"jobType":164,"listed":163,"location":164,"navigation":131,"openRoles":164,"pageLayout":164,"path":1873,"relatedHeading":164,"seo":1874,"series":646,"sitemap":131,"status":164,"stem":1875,"subhead":164,"tags":1876,"video":164,"whyJoin":164,"workplaceType":164,"__hash__":1879},"content\u002Fblog\u002Floop-vs-workflow-vs-agent.md","Loop vs Workflow vs Agent",[1174],{"name":678,"to":196},{"label":198},{"type":166,"value":1177,"toc":1845},[1178,1186,1189,1252,1259,1263,1270,1273,1302,1305,1308,1314,1325,1329,1338,1341,1370,1382,1388,1393,1400,1404,1416,1419,1448,1456,1461,1469,1473,1577,1589,1594,1598,1601,1618,1621,1624,1664,1670,1672,1678,1684,1694,1704,1710,1721,1725,1728,1742,1786,1792,1795,1802,1816,1818,1832,1841],[202,1179,1180,1181,1185],{},"A workflow enumerates steps a person invented. A loop is a compiled standing order that fires when a signal repeats. An agent reasons when the path is not yet known. ",[216,1182,1184],{"href":1183},"https:\u002F\u002Fwww.anthropic.com\u002Fengineering\u002Fbuilding-effective-agents","Anthropic’s guidance on building effective agents"," is blunt about the distinction: encode the job when you can; reserve autonomy for inputs that will not sit still. Operators need the noun that matches how stable the job is — not the noun that wins the demo.",[202,1187,1188],{},"Three words, disambiguated up front, plus the sibling people keep folding in:",[238,1190,1191,1210,1221,1238],{},[241,1192,1193,1196,1197,1200,1201,472,1205,1209],{},[209,1194,1195],{},"Loop."," A compiled ",[209,1198,1199],{},"standing order",": triggers, a trusted recipe, outcomes on a run page, skip when nothing changed. See ",[216,1202,1204],{"href":1203},"\u002Fblog\u002Fwhat-is-loop-engineering\u002F","what is loop engineering",[216,1206,1208],{"href":1207},"\u002Fblog\u002Fsix-things-that-start-a-loop\u002F","six things that start a loop",". Loops are for work that repeats on recognizable signals.",[241,1211,1212,1215,1216,1220],{},[209,1213,1214],{},"Eval loop."," An independent verification that a job finished — tests, schema, read-back, signer — not the model’s claim. See ",[216,1217,1219],{"href":1218},"\u002Fblog\u002Feval-loops-for-enterprise-agent-harnesses\u002F","eval loops for enterprise agent harnesses",". A loop may include eval steps; “eval loop” is not “loop.”",[241,1222,1223,1226,1227,1230,1231,1235,1236,533],{},[209,1224,1225],{},"Agentic workflow."," A ",[209,1228,1229],{},"designed sequence"," of steps toward a goal, with business stop conditions — including a person before a live system changes. See ",[216,1232,1234],{"href":1233},"\u002Fblog\u002Fwhat-is-an-agentic-workflow\u002F","what is an agentic workflow",". Workflows are what someone invented; they are versioned and hosted on a ",[216,1237,948],{"href":947},[241,1239,1240,1243,1244,472,1248,533],{},[209,1241,1242],{},"Agent."," A reasoner that picks tools when the path is not fully enumerated. See ",[216,1245,1247],{"href":1246},"\u002Fblog\u002Fwhat-is-an-agent-harness\u002F","what is an agent harness",[216,1249,1251],{"href":1250},"\u002Fblog\u002Fa-loop-is-not-an-agent\u002F","a loop is not an agent",[202,1253,1254,1255,1258],{},"If your buying conversation starts with “we need an agent,” pause. That sentence often means three different jobs. Buying the wrong one wastes the quarter — not because the model is weak, but because the operating object is wrong. ",[216,1256,1257],{"href":273},"McKinsey’s 2025 State of AI"," keeps showing high usage and uneven scale. Scale, in this vocabulary, usually means compiling what worked — into a workflow or a loop — not hiring a larger model to improvise the same close every month.",[222,1260,1262],{"id":1261},"workflow-steps-a-person-invented","Workflow — steps a person invented",[202,1264,1265,1266,1269],{},"A ",[209,1267,1268],{},"workflow"," is a sequence someone designed: extract, compare to a playbook, quote the CRM fields, wait for the named signer, write or refuse. The author is a person. The instance is durable. The finish line is named before the run starts.",[202,1271,1272],{},"Workflows excel when:",[238,1274,1275,1282,1288,1295],{},[241,1276,1277,1278,1281],{},"the ",[209,1279,1280],{},"stop conditions"," matter as much as the steps — budget, approval, empty result, a clause that is non-standard",[241,1283,1284,1287],{},[209,1285,1286],{},"roles"," differ by step — analyst drafts, controller signs, counsel refuses",[241,1289,1290,1291,1294],{},"the job is ",[209,1292,1293],{},"interactive"," — humans join mid-flight with context attached",[241,1296,1297,1298,1301],{},"you must ",[209,1299,1300],{},"version"," the definition — “which workflow ran last Tuesday?” is a fair audit question",[202,1303,1304],{},"A workflow is the right object when the path is known enough to draw and still needs a person at a gate. Month-end sign-off is a workflow even if a loop assembled the pack. Discount exceptions are a workflow even if a loop watched the stage change. The designed sequence is how you keep roles from collapsing into “the model posted it.”",[202,1306,1307],{},"Workflows fail when teams treat them as magic. A checklist in a prompt is not a workflow. A chat with tools is not a workflow. A Lucidchart on a wiki is not a workflow runtime. A workflow has a durable instance, a finish line, and a record that survives the session. If you hide the chat and cannot reconstruct which step was waiting on a person, you bought conversation.",[202,1309,1310,1313],{},[209,1311,1312],{},"Operator test:"," Can an independent reader reconstruct which steps ran, what was waiting on a person, and what was refused — without Slack search?",[202,1315,1316,1320,1321,1324],{},[216,1317,1319],{"href":1318},"\u002Fblog\u002Fwhat-is-human-in-the-loop-ai\u002F","What is human-in-the-loop AI"," belongs here more than it belongs on the loop. Loops can pause; workflows are often ",[696,1322,1323],{},"about"," the pause. The pause is the product. If your vendor cannot show a named stop, you do not have a workflow. You have a script with optimism.",[222,1326,1328],{"id":1327},"loop-a-compiled-standing-order","Loop — a compiled standing order",[202,1330,1265,1331,1334,1335,1337],{},[209,1332,1333],{},"loop"," is what you run when the world sends the same signal again: Monday close, file in folder, stage flipped, billing event, someone presses run now. See ",[216,1336,1208],{"href":1207},". The author compiled the recipe once. The trigger admits a run. The run page is the contract.",[202,1339,1340],{},"Loops excel when:",[238,1342,1343,1350,1356,1363],{},[241,1344,1345,1346,1349],{},"the recipe is ",[209,1347,1348],{},"stable"," enough to reuse without re-prompting",[241,1351,1352,1355],{},[209,1353,1354],{},"skip semantics"," matter — “nothing changed” should be a first-class outcome",[241,1357,1358,1359,1362],{},"operators need a ",[209,1360,1361],{},"run page",", not a transcript",[241,1364,1365,1366,1369],{},"the same kernel should answer to ",[209,1367,1368],{},"several triggers"," without several copies",[202,1371,1372,1373,1377,1378,1381],{},"Loops are not “set and forget” if forget means no record. A good loop logs quiet outcomes, notifies the roster, and attaches outputs where the next role expects them. ",[216,1374,1376],{"href":1375},"\u002Fblog\u002Fhow-to-evaluate-loop-engineering\u002F","How to evaluate loop engineering"," is the RFP sheet for this object. ",[216,1379,1380],{"href":258},"NIST’s AI Risk Management Framework"," Measure function assumes you can observe those outcomes. A loop without a run page is not observable. It is email archaeology.",[202,1383,1384,1385,1387],{},"Loops fail when they are copied from someone’s channel, when skip is implemented as “don’t email,” or when every step secretly calls a model. On the happy path, ",[216,1386,1251],{"href":1250},": it should not re-derive the job. If Monday’s close still begins with “please do the usual,” you have not compiled anything.",[202,1389,1390,1392],{},[209,1391,1312],{}," Run it ten times with empty inputs. Do you get ten run pages that say skipped — or ten emails saying error?",[202,1394,1395,1399],{},[216,1396,1398],{"href":1397},"\u002Fblog\u002Floop-engineering-vs-harness-engineering\u002F","Loop engineering vs harness engineering"," draws the line: loop engineering compiles repeat work; harness engineering tightens the environment when interactive agents fail. You need both crafts. They fail differently. A loop that guesses at 3am is a harness problem you imported into unattended hours.",[222,1401,1403],{"id":1402},"agent-reasons-when-the-path-is-unknown","Agent — reasons when the path is unknown",[202,1405,1406,1407,1410,1411,1415],{},"An ",[209,1408,1409],{},"agent"," (in the product sense) chooses actions when you cannot enumerate every branch upfront: read these ten documents and tell me which clause conflicts; investigate why margin diverged across three regions; draft three options for a novel pricing exception. ",[216,1412,1414],{"href":1413},"https:\u002F\u002Fdocs.langchain.com\u002Foss\u002Fpython\u002Flangchain\u002Fagents","LangChain’s definition"," is a useful shorthand: Agent = Model + Harness. The harness is not optional. Without tools, stops, and a budget, you have a chatbot with aspirations.",[202,1417,1418],{},"Agents excel when:",[238,1420,1421,1428,1434,1441],{},[241,1422,1423,1424,1427],{},"inputs are ",[209,1425,1426],{},"messy or novel"," — PDFs, threads, ambiguous tickets, a vendor paper the team has not seen",[241,1429,1277,1430,1433],{},[209,1431,1432],{},"path emerges"," during the job — you would not trust a fixed script yet",[241,1435,1436,1437,1440],{},"the cost of a wrong step is ",[209,1438,1439],{},"bounded"," — read-only tools, draft-only outputs",[241,1442,1443,1444,1447],{},"a human will ",[209,1445,1446],{},"compress"," the result into a decision artefact",[202,1449,1450,1451,1455],{},"Agents fail when teams use them as cron with charisma. If the job is the same every Monday, an agent rediscovering the steps is expensive theatre. It is also a control failure: the path that ran last week is not a path you can open next week. ",[216,1452,1454],{"href":1453},"https:\u002F\u002Faiindex.stanford.edu\u002F","Stanford HAI’s AI Index"," is a reminder that capability is not the scarce resource. Operational maturity is.",[202,1457,1458,1460],{},[209,1459,1312],{}," If you hid the chat and kept only the artefact, would the company still know what changed? If not, you bought conversation, not control.",[202,1462,1463,1464,1468],{},"Score agents with ",[216,1465,1467],{"href":1466},"\u002Fblog\u002Fhow-to-evaluate-an-agent-harness\u002F","how to evaluate an agent harness",", not with this page’s skip-semantics tests. The hero metric for an agent is whether you can stop a write and replay what was refused. The hero metric for a loop is whether you can skip quietly and prove it.",[222,1470,1472],{"id":1471},"three-way-comparison","Three-way comparison",[331,1474,1475,1491],{},[334,1476,1477],{},[337,1478,1479,1482,1485,1488],{},[340,1480,1481],{},"Question",[340,1483,1484],{},"Workflow",[340,1486,1487],{},"Loop",[340,1489,1490],{},"Agent",[350,1492,1493,1507,1521,1535,1549,1563],{},[337,1494,1495,1498,1501,1504],{},[355,1496,1497],{},"Who designed the path?",[355,1499,1500],{},"A person encoded steps",[355,1502,1503],{},"A compiled recipe",[355,1505,1506],{},"The model plans within bounds",[337,1508,1509,1512,1515,1518],{},[355,1510,1511],{},"Typical host",[355,1513,1514],{},"Workstream with gates",[355,1516,1517],{},"Recipe library + run page",[355,1519,1520],{},"Workstream or task room",[337,1522,1523,1526,1529,1532],{},[355,1524,1525],{},"Best when",[355,1527,1528],{},"Stops and roles matter",[355,1530,1531],{},"Signal repeats",[355,1533,1534],{},"Path unknown",[337,1536,1537,1540,1543,1546],{},[355,1538,1539],{},"Quiet success",[355,1541,1542],{},"Waiting on signer",[355,1544,1545],{},"Skipped — no change",[355,1547,1548],{},"“Nothing found” with sources",[337,1550,1551,1554,1557,1560],{},[355,1552,1553],{},"Versioning",[355,1555,1556],{},"Workflow definition",[355,1558,1559],{},"Recipe version",[355,1561,1562],{},"Prompt + tool grants",[337,1564,1565,1568,1571,1574],{},[355,1566,1567],{},"Audit reader asks",[355,1569,1570],{},"Which step, which gate",[355,1572,1573],{},"Which trigger, which run",[355,1575,1576],{},"Which sources, which draft",[202,1578,1579,1580,1583,1584,1588],{},"None of these rows replaces your CRM. They describe how work runs ",[209,1581,1582],{},"on top of"," systems of record. The ",[216,1585,1587],{"href":1586},"\u002Fblog\u002Fwhat-is-an-enterprise-ai-operating-system\u002F","enterprise AI operating system"," metaphor is the kernel version of the same problem — isolation, permissions, durable state — applied to jobs rather than applications. This article stays at the operator nouns. Use the OS page when you need the kernel language; use this page when someone says “agent” and means a Monday report.",[202,1590,1591,1593],{},[216,1592,476],{"href":475}," will not tell you which noun to buy. It will ask whether you can name the system, the owner, and the record. Workflows, loops, and agents all fail that test when they live only in a transcript.",[222,1595,1597],{"id":1596},"how-they-compose-in-one-quarter","How they compose in one quarter",[202,1599,1600],{},"Real companies stack all three. The stack is not three products competing. It is three stability levels: novel → designed → repeated.",[419,1602,1603,1608,1613],{},[241,1604,1605,1607],{},[209,1606,1490],{}," investigates a new vendor contract — extracts obligations, flags non-standard terms, drafts a summary for counsel. The path is unknown. The output is a draft. Nothing customer-facing sends.",[241,1609,1610,1612],{},[209,1611,1484],{}," routes the summary through Legal review on a workstream, with a hard stop before anything leaves the company. Roles stay distinct. The instance is durable. Counsel can refuse.",[241,1614,1615,1617],{},[209,1616,1487],{}," watches the signed folder and, every time a countersigned PDF lands, updates the renewal tracker and notifies RevOps — skipping if the hash matches last week’s file. The signal repeats. The recipe does not.",[202,1619,1620],{},"That composition is how you avoid the two failure modes that dominate programmes: agent-for-everything, and checklist-in-a-wiki. The first burns tokens and produces untraceable month six. The second produces a drawing that nobody runs.",[202,1622,1623],{},"Department guides use the same stack with different default nouns:",[238,1625,1626,1640,1654],{},[241,1627,1628,1631,1632,472,1636,533],{},[209,1629,1630],{},"Finance and planning"," — loops for scheduled packs; workflows for close sign-offs; agents for one-off variance forensics. See ",[216,1633,1635],{"href":1634},"\u002Fblog\u002Floops-for-finance-and-planning\u002F","loops for finance and planning",[216,1637,1639],{"href":1638},"\u002Fblog\u002Fcollaborative-ai-for-finance-and-planning\u002F","collaborative AI for finance and planning",[241,1641,1642,1645,1646,472,1650,533],{},[209,1643,1644],{},"Revenue operations"," — loops on stage changes; workflows for discount exceptions; agents for messy account research. See ",[216,1647,1649],{"href":1648},"\u002Fblog\u002Floops-for-revenue-operations\u002F","loops for revenue operations",[216,1651,1653],{"href":1652},"\u002Fblog\u002Fcollaborative-ai-for-revenue-operations\u002F","collaborative AI for revenue operations",[241,1655,1656,1659,1660,533],{},[209,1657,1658],{},"Legal and compliance"," — workflows for approvals; loops on inbound redlines; agents for first-pass clause comparison. See ",[216,1661,1663],{"href":1662},"\u002Fblog\u002Floops-for-legal-and-compliance\u002F","loops for legal and compliance",[202,1665,1666,1669],{},[216,1667,1668],{"href":526},"Collaborative AI"," is the roster pattern underneath all three: multiple roles, one job object, a finish line. The noun you pick does not replace the roster. An agent without a roster is a personal copilot. A loop without a roster notifies a dead channel. A workflow without a roster has nobody to stand at the gate.",[222,1671,897],{"id":896},[202,1673,1674,1677],{},[209,1675,1676],{},"Agent for everything."," Impressive week one; untraceable month six. The model will happily re-solve a solved problem. You will pay for it twice: inference, and the incident when it improvises a write.",[202,1679,1680,1683],{},[209,1681,1682],{},"Workflow without an instance."," A diagram is not a runtime. If you cannot open last Tuesday’s run and see the step that was waiting, you have documentation, not a workflow.",[202,1685,1686,1689,1690,1693],{},[209,1687,1688],{},"Loop copied from someone’s Slack channel."," If reuse means “find the old thread,” you have folklore, not ",[216,1691,1692],{"href":1203},"loop engineering",". Folklore does not survive vacation.",[202,1695,1696,1699,1700,1703],{},[209,1697,1698],{},"Confusing eval loops with loops."," Benchmarks and read-backs verify completion; standing-order loops start and record repeat work. Both can coexist on one job. Collapsing them hides whether you can start work ",[696,1701,1702],{},"and"," whether you can prove it finished.",[202,1705,1706,1709],{},[209,1707,1708],{},"Buying a platform that only sells one noun."," If the vendor cannot host a workflow gate, a loop run page, and an agent task on the same roster, you will rebuild org boundaries in email. That is not a tooling preference. It is how departments already work.",[202,1711,1712,1715,1716,1720],{},[209,1713,1714],{},"Treating RPA as the fourth synonym."," Screen replay is a different category. See ",[216,1717,1719],{"href":1718},"\u002Fblog\u002Floop-vs-rpa\u002F","loop vs RPA",". A bot that clicks an ERP has a place. It is not a standing-order loop, a designed workflow, or a reasoner.",[222,1722,1724],{"id":1723},"choosing-this-week","Choosing this week",[202,1726,1727],{},"Ask two questions, in this order:",[419,1729,1730,1736],{},[241,1731,1732,1735],{},[209,1733,1734],{},"Do we know the steps?"," No → agent, bounded. Yes → workflow or loop.",[241,1737,1738,1741],{},[209,1739,1740],{},"Will the same signal fire again?"," Yes → loop. No → workflow.",[1743,1744,1748],"pre",{"className":1745,"code":1746,"language":1747,"meta":169,"style":169},"language-mermaid shiki shiki-themes github-light github-dark","flowchart TD\n  steps{\"Do we know the steps?\"}\n  steps -->|no| agentPick[\"Agent: explore, keep a draft\"]\n  steps -->|yes| signal{\"Will the same signal fire again?\"}\n  signal -->|yes| loopPick[\"Loop: standing order\"]\n  signal -->|no| workflowPick[\"Workflow: designed sequence with stops\"]\n","mermaid",[762,1749,1750,1758,1763,1768,1774,1780],{"__ignoreMap":169},[1751,1752,1755],"span",{"class":1753,"line":1754},"line",1,[1751,1756,1757],{},"flowchart TD\n",[1751,1759,1760],{"class":1753,"line":170},[1751,1761,1762],{},"  steps{\"Do we know the steps?\"}\n",[1751,1764,1765],{"class":1753,"line":1128},[1751,1766,1767],{},"  steps -->|no| agentPick[\"Agent: explore, keep a draft\"]\n",[1751,1769,1771],{"class":1753,"line":1770},4,[1751,1772,1773],{},"  steps -->|yes| signal{\"Will the same signal fire again?\"}\n",[1751,1775,1777],{"class":1753,"line":1776},5,[1751,1778,1779],{},"  signal -->|yes| loopPick[\"Loop: standing order\"]\n",[1751,1781,1783],{"class":1753,"line":1782},6,[1751,1784,1785],{},"  signal -->|no| workflowPick[\"Workflow: designed sequence with stops\"]\n",[202,1787,1788,1789,1791],{},"If both yes — stable steps ",[696,1790,1702],{}," a repeating signal — start with a loop. Add workflow gates inside it when a named signer must appear. Add an agent step when one stage still needs reading messy inputs. That is the composition above, reduced to a decision tree you can use in a working session.",[202,1793,1794],{},"If both no — unknown steps and a one-off — you are in research. Do not compile yet. Do not buy a bot. Bound the tools, keep the output as a draft, and decide next week whether the path is now known.",[202,1796,1797,1801],{},[216,1798,1800],{"href":1799},"\u002Fblog\u002Ffour-pillars-of-an-enterprise-ai-platform\u002F","The four pillars of an enterprise AI platform"," is the wider map: communication brings context, collaboration hosts the job, automate repeats the recipe, governance refuses what should not land. Loops, workflows, and agents all sit on that map. They are not a substitute for it.",[202,1803,1804,1805,1098,1807,1101,1811,1815],{},"Related reading: ",[216,1806,527],{"href":526},[216,1808,1810],{"href":1809},"\u002Fblog\u002Fhow-to-evaluate-collaborative-ai\u002F","how to evaluate collaborative AI",[216,1812,1814],{"href":1813},"\u002Fblog\u002Fwhat-is-harness-engineering\u002F","what is harness engineering"," when agents and loops share the same stops.",[222,1817,506],{"id":505},[202,1819,1820,1821,1823,1824,1827,1828,1831],{},"Nimbus hosts all three nouns on the same ",[216,1822,948],{"href":977},". Loops are standing orders with run pages. Agentic workflows are designed sequences with ",[216,1825,1826],{"href":512},"governance"," gates. Agents run as bounded teammates with scoped grants, not as a second platform. ",[216,1829,15],{"href":1830},"\u002Fproduct\u002Fconflux\u002F"," is often the communication door that starts a loop when a file lands.",[202,1833,1834,1835,472,1838,1840],{},"Treat that as one vendor’s mapping of the pattern, not as the definition. If you are writing an RFP, paste the questions from ",[216,1836,1837],{"href":1375},"how to evaluate loop engineering",[216,1839,1467],{"href":1466}," and run them on whoever claims the nouns. The useful outcome is not a logo. It is whether Tuesday’s close, Tuesday’s exception, and Tuesday’s novel contract can share a roster without collapsing into one chat.",[1842,1843,1844],"style",{},"html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}html.dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}",{"title":169,"searchDepth":170,"depth":170,"links":1846},[1847,1848,1849,1850,1851,1852,1853,1854],{"id":1261,"depth":170,"text":1262},{"id":1327,"depth":170,"text":1328},{"id":1402,"depth":170,"text":1403},{"id":1471,"depth":170,"text":1472},{"id":1596,"depth":170,"text":1597},{"id":896,"depth":170,"text":897},{"id":1723,"depth":170,"text":1724},{"id":505,"depth":170,"text":506},"2026-09-12","A workflow enumerates steps a person invented.",{"eyebrow":1858,"title":1859},"Common questions","Standing orders, sequences, and reasoners",[1861,1864,1867,1870],{"question":1862,"answer":1863},"Can a loop contain an agent?","Yes, and that composition is often the adult design. A loop is the standing order — triggers, recipe, run page, skip semantics. A step inside may call an agent when the input is messy, the same way a close pack may include a first-pass read of a PDF. The loop still owns notifications, quiet outcomes, and where artefacts land. Use this pattern when one stage is genuinely unstructured and the rest is known. Refuse a design that lets the agent rewrite the recipe, skip the signer, or hide its tool calls off the run page. The reasoner is a guest on the standing order, not the owner of the job.",{"question":1865,"answer":1866},"Is an agentic workflow just a loop with extra steps?","Not quite, and the difference shows up under audit. Workflows are authored sequences with explicit stops — often interactive, versioned per department, and meant for jobs that need a person mid-flight. Loops are reused recipes optimized for repeat signals and quiet skips. Many teams use both on the same workstream: the loop watches the folder, the workflow carries the exception that needs counsel. Choose a workflow when stop conditions and roles matter as much as the steps. Choose a loop when the same signal will fire again and nobody should re-describe the job. Refuse a vendor that uses the two words interchangeably on slide one and cannot show both objects in a demo.",{"question":1868,"answer":1869},"When should I choose an agent instead of a loop?","When nobody can write the steps yet — novel research, one-off negotiation, exploratory analysis, a contract shape the team has not seen. Agents earn their keep on the unknown path. Once the path stabilizes, compile it into a loop or a workflow so the company is not paying for rediscovery every Monday. Use an agent when the cost of a wrong step is bounded — read-only tools, draft-only outputs, a human who will compress the result. Refuse an agent as a Monday cron job. That is expensive theatre, and it fails the first time the model improvises a write. See [a loop is not an agent](\u002Fblog\u002Fa-loop-is-not-an-agent\u002F).",{"question":1871,"answer":1872},"How do I keep the three nouns from collapsing in an RFP?","Score them on different sheets. Ask for a designed sequence with a durable instance and a named stop — that is the workflow. Ask for a repeating signal, a skipped run, and a recipe you can clone — that is the loop. Ask for a bounded reasoner that cites sources and cannot write without a signer — that is the agent. [How to evaluate loop engineering](\u002Fblog\u002Fhow-to-evaluate-loop-engineering\u002F) and [how to evaluate an agent harness](\u002Fblog\u002Fhow-to-evaluate-an-agent-harness\u002F) exist because one checklist cannot cover both. Refuse a single “agentic platform” scorecard that starts with context-window size. That sheet buys a model for a job that needed a standing order.","\u002Fblog\u002Floop-vs-workflow-vs-agent",{"title":1172,"description":1856},"blog\u002Floop-vs-workflow-vs-agent",[1877,646,1878,1167],"loops","agentic-workflow","wur-D-OIn6LTECy7powb31UfYHPvIHJ1J9zX_f1SJdM",{"fold":1881,"id":1885,"title":1886,"archived":163,"authors":164,"badge":164,"body":1887,"date":164,"definedTerm":164,"department":164,"description":169,"extension":172,"eyebrow":164,"faqHeader":164,"faqs":164,"footerBand":1891,"headline":164,"image":164,"industry":164,"jobType":164,"listed":131,"location":164,"navigation":131,"openRoles":164,"pageLayout":164,"path":1895,"relatedHeading":164,"seo":1896,"series":164,"sitemap":163,"status":164,"stem":1897,"subhead":164,"tags":164,"video":164,"whyJoin":164,"workplaceType":164,"__hash__":1898},{"headline":1882,"description":1883,"primaryLabel":8,"primaryTo":1884,"secondaryLabel":667,"secondaryTo":12},"Run frontier AI your business actually owns.","Governed workstreams, 3,000+ integrations, and a proprietary knowledge graph. Start on Free.","\u002Fsignup?plan=free","content\u002Fshared\u002Fcta.md","Site CTAs",{"type":166,"value":1888,"toc":1889},[],{"title":169,"searchDepth":170,"depth":170,"links":1890},[],{"headline":1892,"description":1893,"primaryLabel":8,"primaryTo":1884,"secondaryLabel":1894,"secondaryTo":85},"See what governed AI looks like on your stack.","Connect your tools, run a workstream, and keep every decision on your ledger. Start on Free.","Talk to our team","\u002Fshared\u002Fcta",{"title":1886,"description":169},"shared\u002Fcta","eYqahyaPnbp8GKrWpoORbZdtmkWmgHr5F61ZHOnb8sY",1792999478305]