[{"data":1,"prerenderedAt":1787},["ShallowReactive",2],{"site-nav-content":3,"blog:\u002Fblog\u002Fwhat-is-an-ai-workstream":177,"blog-index-copy":688,"blog:\u002Fblog\u002Fwhat-is-an-ai-workstream:surround":709,"hiring-banner-content":1756,"site-cta-content":1768},{"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",{"id":178,"title":179,"archived":163,"authors":180,"badge":184,"body":186,"date":664,"definedTerm":164,"department":164,"description":665,"extension":172,"eyebrow":164,"faqHeader":666,"faqs":669,"footerBand":164,"headline":164,"image":164,"industry":164,"jobType":164,"listed":131,"location":164,"navigation":131,"openRoles":164,"pageLayout":164,"path":679,"relatedHeading":164,"seo":680,"series":681,"sitemap":131,"status":164,"stem":682,"subhead":164,"tags":683,"video":164,"whyJoin":164,"workplaceType":164,"__hash__":687},"content\u002Fblog\u002Fwhat-is-an-ai-workstream.md","What is an AI Workstream",[181],{"name":182,"to":183},"Nimbus Research","https:\u002F\u002Fgonimbus.ai",{"label":185},"Explainer",{"type":166,"value":187,"toc":639},[188,197,200,203,212,215,220,305,313,317,323,326,345,348,373,376,381,387,396,402,408,417,421,427,432,438,444,450,456,459,467,471,478,481,488,507,511,515,518,522,525,529,532,536,542,546,551,555,558,562,569,573,576,580,587,591,594,598,601,605,610,614,623,627],[189,190,191,192,196],"p",{},"An AI workstream is a ",[193,194,195],"strong",{},"shared workspace for one job",": a brief, the tools allowed, the people and AI on it, a budget, and a finish line.",[189,198,199],{},"A Slack channel with a bot is not a job. It is a room. Anyone can paste anything. The bot never knows when the work is done. Next quarter, nobody can say which systems were in play or who was allowed to change them.",[189,201,202],{},"If you cannot name the systems in scope and the approval policy on writes, you do not have a workstream. You have a conversation.",[189,204,205,206,211],{},"Software teams already learned this. Work lives in issues and tickets, not in unbounded chat. ",[207,208,210],"a",{"href":209},"https:\u002F\u002Fwww.atlassian.com\u002Fagile\u002Fproject-management\u002Fepics-stories-themes","Atlassian’s epics and stories"," are named packages with a boundary. AI operations are still catching up. The missing object is often the work package: a place where the job actually lives.",[189,213,214],{},"The analogy is not decoration. Tickets have a requester, a scope, an owner, and a closed state. Copilots have a thread. Threads do not archive cleanly, do not attach least-privilege connectors, and do not carry a named signer. When AI started touching live systems, the thread stopped being a sufficient container.",[216,217,219],"h2",{"id":218},"words-youll-hear","Words you’ll hear",[221,222,223,230,236,252,258,269,275,285,295],"ul",{},[224,225,226,229],"li",{},[193,227,228],{},"Brief."," What this job is for, and what “done” means. At work, “Q3 regional discount hygiene” is a brief. “My stuff” is not.",[224,231,232,235],{},[193,233,234],{},"Connector."," A secure link to a live system (CRM, ERP, Drive). Attach what this job needs — not every system “just in case.” Default is read-only.",[224,237,238,241,242,246,247,251],{},[193,239,240],{},"Scope \u002F least privilege."," Only the data and tools required for ",[243,244,245],"em",{},"this"," job. ",[207,248,250],{"href":249},"https:\u002F\u002Fwww.hhs.gov\u002Fhipaa\u002Ffor-professionals\u002Fprivacy\u002Fguidance\u002Fminimum-necessary-requirement\u002Findex.html","HIPAA’s minimum necessary"," is the same instinct: do not attach every system to every task. GDPR purpose limitation is the privacy-law cousin.",[224,253,254,257],{},[193,255,256],{},"God workspace."," One org-wide chat that can see every folder and every CRM object because setup was easier. At work, this is how recruiting sees finance forecasts.",[224,259,260,263,264,268],{},[193,261,262],{},"Agent team."," The AI specialists assigned to the job. The workstream is the stage; the team is the cast. See ",[207,265,267],{"href":266},"\u002Fblog\u002Fwhat-is-multi-agent-ai\u002F","What is multi-agent AI",".",[224,270,271,274],{},[193,272,273],{},"Named signer."," Who must approve a write. At work, this is a role that already owns that class of change.",[224,276,277,280,281,268],{},[193,278,279],{},"Budget."," The spend ceiling for the job. See ",[207,282,284],{"href":283},"\u002Fblog\u002Fwhat-is-ai-token-economics\u002F","What is AI token economics",[224,286,287,290,291,268],{},[193,288,289],{},"Wiki section."," The asserted playbooks this job may load. See ",[207,292,294],{"href":293},"\u002Fblog\u002Fwhat-is-a-company-wiki-for-ai-agents\u002F","What is a company wiki for AI agents",[224,296,297,300,301,268],{},[193,298,299],{},"Lifecycle Graph."," The chain this job emits as it runs. See ",[207,302,304],{"href":303},"\u002Fblog\u002Fwhat-is-a-lifecycle-graph\u002F","What is a lifecycle graph",[189,306,307,308,312],{},"An ",[207,309,311],{"href":310},"\u002Fblog\u002Fwhat-is-an-agentic-workflow\u002F","agentic workflow"," is the sequence of steps. The workstream is the durable instance those steps run inside. A workflow definition without a workstream is a script on someone’s laptop. A workstream without a workflow is a folder with no process.",[216,314,316],{"id":315},"why-you-should-care","Why you should care",[189,318,319,320,322],{},"Without a boundary, two departments sharing an AI tool will either over-share (the recruiting job can see finance forecasts) or under-share (people export spreadsheets to personal ChatGPT). The workstream is the compromise: enough context to do ",[243,321,245],{}," job, not the whole company.",[189,324,325],{},"It affects you if work:",[221,327,328,331,334,337,340],{},[224,329,330],{},"touches more than one system",[224,332,333],{},"involves more than one role",[224,335,336],{},"can change a live record",[224,338,339],{},"needs a budget you can attribute",[224,341,342,343],{},"must still be explainable after people leave — see ",[207,344,304],{"href":303},[189,346,347],{},"Open workstreams the way you would open a ticket:",[221,349,350,357,364,367,370],{},[224,351,352,353,356],{},"One workstream per ",[193,354,355],{},"outcome",", not per person. “Q3 regional discount hygiene” can have several humans. “My stuff” cannot be governed or archived.",[224,358,359,360,363],{},"Attach the ",[193,361,362],{},"minimum"," connectors.",[224,365,366],{},"Set the write policy on day one, even if you start read-only.",[224,368,369],{},"Reuse templates, not last month’s chat thread.",[224,371,372],{},"Close or archive when the job ends. A sprint that never ends is not a sprint.",[189,374,375],{},"A standing “Ask AI” workstream with org-wide connectors recreates the copilot, including the blast radius.",[377,378,380],"h3",{"id":379},"what-changes-by-role","What changes by role",[189,382,383,386],{},[193,384,385],{},"Finance."," Chargeback becomes possible because the job is named. Close workstreams can attach ERP read-only, load the close checklist from the wiki, and keep GTM out of the ledger. A company-wide AI pool with no workstream attribution is a shared kitchen.",[189,388,389,392,393,395],{},[193,390,391],{},"Legal."," Scope is a processing purpose. A workstream for a renewal can include legal and go-to-market on ",[243,394,245],{}," goal without merging their entire universes. Legal also gets a closed state: when the job ends, retention follows the type of record instead of an immortal channel.",[189,397,398,401],{},[193,399,400],{},"Operations."," This is the ticket analogue they already wanted. Ops should refuse god workspaces, insist on a finish line, and treat human wait as a status, not a side conversation in Slack.",[189,403,404,407],{},[193,405,406],{},"Go-to-market."," Cross-functional launches finally have a place that is not a merged Slack. GTM still should not get finance’s ERP “for context.” Templates beat copying last quarter’s thread, which silently copies last quarter’s over-attached connectors.",[189,409,410,413,414,416],{},[193,411,412],{},"Security."," Least privilege is now a product object, not a memo. Connectors default to read-only. Adding a write path is a deliberate change to ",[243,415,245],{}," job, not a tenant-wide toggle. A workstream that never closes is a standing access grant.",[377,418,420],{"id":419},"what-people-get-wrong","What people get wrong",[189,422,423,426],{},[193,424,425],{},"One workstream per person."," You cannot archive “my stuff.” You cannot attribute it. You cannot apply least privilege.",[189,428,429,431],{},[193,430,256],{}," Setup is easier. Blast radius is the company.",[189,433,434,437],{},[193,435,436],{},"ChatGPT Projecthe unit."," Some files, some instructions. Typically no connector-level least privilege, quoted writes, spend caps, or lasting record. Fine for personal research. Not an operations unit.",[189,439,440,443],{},[193,441,442],{},"Too small."," If setup exceeds the job, use a lighter sanctioned copilot path. Do not open a workstream to rewrite one sentence.",[189,445,446,449],{},[193,447,448],{},"Too large."," If you cannot explain the purpose in one sentence, or you keep attaching “one more connector,” split.",[189,451,452,455],{},[193,453,454],{},"Never closing."," Standing rooms recreate Slack, including the archaeology problem.",[189,457,458],{},"Good looks like: one outcome, minimum connectors, write policy on day one, wiki sections subscribed, budget capped, named signer, archive when done. Failure looks like an org-wide copilot with every OAuth grant and a channel that outlives the campaign.",[189,460,461,462,466],{},"The ",[207,463,465],{"href":464},"\u002Fblog\u002Fwhat-is-an-enterprise-ai-operating-system\u002F","enterprise AI OS"," metaphor is isolation plus I\u002FO plus state. The workstream is the isolation unit. Without it, connectors, wiki, and agent teams have nowhere to attach that an auditor could name.",[216,468,470],{"id":469},"how-this-shows-up-in-nimbus","How this shows up in Nimbus",[189,472,473,474,477],{},"In Nimbus, workstreams are how ",[207,475,476],{"href":310},"agentic workflows"," become company objects rather than a file only one engineer can run.",[189,479,480],{},"Each workstream carries a brief, wiki sections (approved playbooks), connector attachments (read-only by default), agent team assignment, spend budget, release policy on writes, and nodes on the Lifecycle Graph.",[189,482,483,484,487],{},"Cross-department work is multiple teams on one workstream, not a merged Slack. Operators open this themselves; the point of an ",[207,485,486],{"href":464},"enterprise AI operating system"," is that the job folder is a product, not a forward-deployed spreadsheet.",[189,489,490,491,494,495,499,500,499,504,268],{},"See ",[207,492,31],{"href":493},"\u002Fproduct\u002Fworkstreams\u002F",". Related product: ",[207,496,498],{"href":497},"\u002Fproduct\u002Fagent-teams\u002F","Agent teams",", ",[207,501,503],{"href":502},"\u002Fproduct\u002Fwiki\u002F","Wiki",[207,505,39],{"href":506},"\u002Fproduct\u002Fgovernance\u002F",[216,508,510],{"id":509},"questions-people-actually-ask","Questions people actually ask",[377,512,514],{"id":513},"is-a-chatgpt-project-a-workstream","Is a ChatGPT “Project” a workstream?",[189,516,517],{},"It is a weak analogue: some files, some custom instructions. It typically lacks connector-level least privilege, quoted writes, spend caps, and a lasting record. Useful for personal research. Not an operations unit.",[377,519,521],{"id":520},"how-small-is-too-small","How small is too small?",[189,523,524],{},"If the setup cost exceeds the job, use a lighter sanctioned copilot path. Do not create a workstream to rewrite one sentence.",[377,526,528],{"id":527},"how-large-is-too-large","How large is too large?",[189,530,531],{},"If you cannot explain the purpose in one sentence, or you keep attaching “one more connector,” split.",[377,533,535],{"id":534},"can-one-workstream-serve-multiple-departments","Can one workstream serve multiple departments?",[189,537,538,539,541],{},"Yes — go-to-market and legal on a renewal, for example. They share ",[243,540,245],{}," goal’s scope, not each other’s entire universe.",[377,543,545],{"id":544},"how-do-we-budget-them","How do we budget them?",[189,547,548,549,268],{},"Caps per workstream, plus an organisation pool. Chargeback by workstream beats “the AI bill.” See ",[207,550,284],{"href":283},[377,552,554],{"id":553},"is-a-slack-channel-with-a-bot-enough-if-we-add-a-approve-command","Is a Slack channel with a bot enough if we add a \u002Fapprove command?",[189,556,557],{},"No. A command is not connector least privilege, a quoted payload, a durable chain, or an archive policy. It is still a room.",[377,559,561],{"id":560},"who-is-allowed-to-open-a-workstream","Who is allowed to open a workstream?",[189,563,564,565,268],{},"Whoever is allowed to open that class of job in analogue life — with the same instinct as who may open a ticket or a change request. An “AI team” bottleneck recreates the waitlist that causes ",[207,566,568],{"href":567},"\u002Fblog\u002Fwhat-is-shadow-ai\u002F","shadow AI",[377,570,572],{"id":571},"what-happens-when-the-job-ends","What happens when the job ends?",[189,574,575],{},"Close or archive. Revoke standing connector usefulness. Keep the reconstructable chain according to retention, not the entire chat.",[377,577,579],{"id":578},"do-we-need-a-workstream-for-read-only-analysis","Do we need a workstream for read-only analysis?",[189,581,582,583,268],{},"When the analysis crosses systems, roles, or must be replayed later, yes. When it is personal drafting with no live-system scope, a sanctioned copilot may be enough. See ",[207,584,586],{"href":585},"\u002Fblog\u002Fcollaborative-ai-and-personal-assistants\u002F","collaborative AI and personal assistants",[377,588,590],{"id":589},"how-do-wiki-and-connectors-differ-inside-a-workstream","How do wiki and connectors differ inside a workstream?",[189,592,593],{},"Wiki is asserted policy the job must follow. Connectors are live systems the job may read (and, if enabled, write). Mixing them into one “knowledge” pile is how Drive folklore overwrites the playbook.",[377,595,597],{"id":596},"can-we-keep-one-standing-workstream-for-ask-anything","Can we keep one standing workstream for “ask anything”?",[189,599,600],{},"You can. You will recreate the copilot, including over-share. Standing Q&A belongs on a tightly scoped, read-only path if it exists at all.",[377,602,604],{"id":603},"how-does-this-relate-to-agent-teams","How does this relate to agent teams?",[189,606,607,608,268],{},"The workstream is the job. The agent team is the cast assigned to it. Changing the cast does not change the brief, the connectors, or the signer. See ",[207,609,267],{"href":266},[216,611,613],{"id":612},"related-reading","Related reading",[189,615,616,619,620,268],{},[207,617,618],{"href":464},"What is an enterprise AI operating system"," and ",[207,621,622],{"href":310},"What is an agentic workflow",[216,624,626],{"id":625},"sources","Sources",[221,628,629,634],{},[224,630,631],{},[207,632,633],{"href":209},"Atlassian, epics, stories, and themes",[224,635,636],{},[207,637,638],{"href":249},"HHS, HIPAA minimum necessary requirement",{"title":169,"searchDepth":170,"depth":170,"links":640},[641,642,647,648,662,663],{"id":218,"depth":170,"text":219},{"id":315,"depth":170,"text":316,"children":643},[644,646],{"id":379,"depth":645,"text":380},3,{"id":419,"depth":645,"text":420},{"id":469,"depth":170,"text":470},{"id":509,"depth":170,"text":510,"children":649},[650,651,652,653,654,655,656,657,658,659,660,661],{"id":513,"depth":645,"text":514},{"id":520,"depth":645,"text":521},{"id":527,"depth":645,"text":528},{"id":534,"depth":645,"text":535},{"id":544,"depth":645,"text":545},{"id":553,"depth":645,"text":554},{"id":560,"depth":645,"text":561},{"id":571,"depth":645,"text":572},{"id":578,"depth":645,"text":579},{"id":589,"depth":645,"text":590},{"id":596,"depth":645,"text":597},{"id":603,"depth":645,"text":604},{"id":612,"depth":170,"text":613},{"id":625,"depth":170,"text":626},"2026-08-17","An AI workstream is one shared job: a brief, the tools allowed, the people and AI on it, a budget, and a finish line.",{"eyebrow":667,"title":668},"Short answers","One job, one shared place",[670,673,676],{"question":671,"answer":672},"What is an AI workstream?","A shared workspace for one job: a brief, the tools allowed, the people and AI on it, a budget, and a finish line.",{"question":674,"answer":675},"Is a Slack channel with a bot a workstream?","No. A channel has no brief, no allowed-tool list, and no finish line a person can refuse.",{"question":677,"answer":678},"What makes it finish?","A named outcome and someone who can say the work does not go out.","\u002Fblog\u002Fwhat-is-an-ai-workstream",{"title":179,"description":665},"explainer","blog\u002Fwhat-is-an-ai-workstream",[681,684,685,686],"workstreams","enterprise-ai","governance","GqysvBAVOydIlMQ2bTUJZvFY1Z9D-lp9m7VqL4XVPA0",{"hero":689,"id":691,"title":692,"archived":163,"authors":164,"badge":164,"body":693,"date":164,"definedTerm":164,"department":164,"description":697,"extension":172,"eyebrow":698,"faqHeader":164,"faqs":164,"footerBand":699,"headline":164,"image":164,"industry":164,"jobType":164,"listed":131,"location":164,"navigation":131,"openRoles":164,"pageLayout":164,"path":60,"relatedHeading":705,"seo":706,"series":164,"sitemap":131,"status":164,"stem":707,"subhead":164,"tags":164,"video":164,"whyJoin":164,"workplaceType":164,"__hash__":708},{"filename":690},"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":694,"toc":695},[],{"title":169,"searchDepth":170,"depth":170,"links":696},[],"Deep dives into pre-cognitive intelligence, sentient enterprises, and the evolving landscape of AI-driven business transformation.","Latest Research",{"headline":700,"description":701,"primaryLabel":702,"primaryTo":703,"secondaryLabel":704,"secondaryTo":12},"Stay at the frontier.","Subscribe for product updates and new insights.","Subscribe","\u002Fnewsletter","Explore the platform","More research",{"title":692,"description":697},"blog\u002Findex","BFSWGYO9bcTlaulivKYWyg08_DJHsdGg3OC6g_CG1Hw",[710,1201],{"id":711,"title":712,"archived":163,"authors":713,"badge":715,"body":716,"date":664,"definedTerm":164,"department":164,"description":720,"extension":172,"eyebrow":164,"faqHeader":1183,"faqs":1185,"footerBand":164,"headline":164,"image":164,"industry":164,"jobType":164,"listed":131,"location":164,"navigation":131,"openRoles":164,"pageLayout":164,"path":1195,"relatedHeading":164,"seo":1196,"series":681,"sitemap":131,"status":164,"stem":1197,"subhead":164,"tags":1198,"video":164,"whyJoin":164,"workplaceType":164,"__hash__":1200},"content\u002Fblog\u002Fwhat-is-an-enterprise-ai-operating-system.md","What is an Enterprise AI Operating System",[714],{"name":182,"to":183},{"label":185},{"type":166,"value":717,"toc":1159},[718,721,724,727,734,737,739,808,811,844,856,863,865,871,874,877,880,883,885,890,903,908,913,918,920,926,940,946,952,958,961,971,973,976,1016,1034,1036,1040,1043,1047,1055,1059,1062,1066,1069,1073,1076,1080,1083,1087,1090,1094,1099,1103,1106,1110,1115,1119,1122,1126,1137,1139,1145,1147],[189,719,720],{},"An enterprise AI operating system is the layer between the AI model and how departments actually work — like Windows sits between the chip and your apps.",[189,722,723],{},"If your question is “which model should we buy,” you are shopping for a chip. If your question is “how do revenue, legal, and finance run the same loop without a personal-account workaround,” you are shopping for an OS.",[189,725,726],{},"It is not a chatbot with company login. It is not a model API with a prompt library.",[189,728,729,733],{},[207,730,732],{"href":731},"https:\u002F\u002Fwww.mckinsey.com\u002Fcapabilities\u002Fquantumblack\u002Four-insights\u002Fthe-state-of-ai","McKinsey’s 2025 State of AI survey"," found that 88% of companies use AI in at least one function — and that a majority are still piloting. About one in three report that they are scaling. Buying another model does not close that gap. The missing layer is how work actually runs.",[189,735,736],{},"The OS metaphor is useful if you keep it honest. An operating system does not replace your spreadsheet or your CRM. It gives applications isolation, permissions, input and output, and a place to keep state after the window closes. An enterprise AI OS does the same for work that uses models: isolation of jobs, rights over tools and data, reads and writes to live systems, budgets, and a record that survives the session.",[216,738,219],{"id":218},[221,740,741,747,753,763,769,780,790,800],{},[224,742,743,746],{},[193,744,745],{},"Copilot."," A high-quality assistant for a person. Admin controls, company login. Not, by itself, how several departments finish one job under a named signer. At work, this is “help me draft.” It is not “release this CRM change.”",[224,748,749,752],{},[193,750,751],{},"Operating system (in this sense)."," Process isolation, permissions, input\u002Foutput to live systems, budgeting, and durable state — the jobs a kernel does for apps.",[224,754,755,758,759,762],{},[193,756,757],{},"System of record."," CRM, ERP, HR — still authoritative. The OS is the system of ",[243,760,761],{},"work",", not a second CRM.",[224,764,765,768],{},[193,766,767],{},"Forward-deployed engineer."," A vendor consultant who sits with you for months. Some programmes need that. Many companies need governed work this quarter without it.",[224,770,771,774,775,779],{},[193,772,773],{},"NIST AI RMF."," ",[207,776,778],{"href":777},"https:\u002F\u002Fwww.nist.gov\u002Fitl\u002Fai-risk-management-framework","Govern, Map, Measure, Manage"," — public-sector language for the same kernel idea: identity, tool rights, and a record attached to real actions.",[224,781,782,785,786,268],{},[193,783,784],{},"Workstream."," The process-isolation unit: one job, one scope, one finish line. See ",[207,787,789],{"href":788},"\u002Fblog\u002Fwhat-is-an-ai-workstream\u002F","What is an AI workstream",[224,791,792,795,796,268],{},[193,793,794],{},"Fail-closed writes."," Missing named signer means nothing happens. See ",[207,797,799],{"href":798},"\u002Fblog\u002Fwhat-is-write-back-governance\u002F","What is write-back governance",[224,801,802,805,806,268],{},[193,803,804],{},"Usage limits and quotes."," Work is quoted before it runs and capped by limits and budgets. See ",[207,807,284],{"href":283},[189,809,810],{},"Five jobs cluster around the term:",[812,813,814,820,826,832,838],"ol",{},[224,815,816,819],{},[193,817,818],{},"Process isolation."," Go-to-market does not silently inherit finance’s ERP login.",[224,821,822,825],{},[193,823,824],{},"Resource management."," Inference and tool calls are budgeted.",[224,827,828,831],{},[193,829,830],{},"I\u002FO control."," Reads and writes to CRM and ERP are first-class — not “chat that sometimes calls an API.”",[224,833,834,837],{},[193,835,836],{},"Permissioning."," Identity and context decide what an agent can see and do. A signed PDF is not enforcement.",[224,839,840,843],{},[193,841,842],{},"Durable state."," Outcomes, approvals, and rationale survive the session.",[189,845,846,847,619,851,855],{},"Copilots generally fail the last three. ",[207,848,850],{"href":849},"\u002Fblog\u002Fnimbus-vs-chatgpt-enterprise\u002F","ChatGPT Enterprise",[207,852,854],{"href":853},"\u002Fblog\u002Fnimbus-vs-claude\u002F","Claude for Work"," are excellent assistants. They are not this job.",[189,857,858,862],{},[207,859,861],{"href":860},"\u002Fblog\u002Fwhat-is-model-context-protocol\u002F","Model Context Protocol"," is also not this job. A common plug for tools is USB. USB did not create Windows.",[216,864,316],{"id":315},[189,866,867,868,870],{},"Operators do not “open the OS” the way they open a model playground. They open ",[193,869,761],{},": a brief, a scoped live system, a review, a release.",[189,872,873],{},"It affects you if AI is starting to touch revenue, financial close, customer records, or regulated processes. Chat history does not answer “who approved this, against which policy?”",[189,875,876],{},"The OS also matters if you refuse a six-to-twelve-month vendor-engineer programme as the only path to production.",[189,878,879],{},"The anti-pattern is using an OS as a better chatbot: one user, one thread, no write path, no memory beyond the conversation. If nobody except the original operator can reconstruct what happened, you have a log, not an operating system.",[189,881,882],{},"Personal copilots optimise for “the model always answers.” An OS optimises for “the company only acts when the gate says so.” A spend cap or a missing approval is a successful outcome.",[377,884,380],{"id":379},[189,886,887,889],{},[193,888,385],{}," The OS is how close and forecast jobs get a budget, a read-only ERP connector, a wiki checklist, and a named signer — without a second ledger. Finance should still own NetSuite. The OS should point at it.",[189,891,892,894,895,898,899,268],{},[193,893,391],{}," Reconstructable authorisation, purpose-limited scope, and a place that is not a personal chat vendor. Legal should evaluate whether unapproved writes are ",[243,896,897],{},"impossible",", not whether a policy PDF exists. See ",[207,900,902],{"href":901},"\u002Fblog\u002Fwhat-is-ai-governance\u002F","What is AI governance",[189,904,905,907],{},[193,906,400],{}," Isolation and durable state are ops problems. Ops should ask whether a paused run is a first-class object, whether connectors default to read-only, and whether Perception (or equivalent) can answer “why did this change?” without a data team reconstructing Slack.",[189,909,910,912],{},[193,911,406],{}," Cross-department loops — legal on a renewal, finance on a discount — need a shared job, not a shared inbox. GTM should not have to choose between a copilot that cannot write safely and a spreadsheet export to a consumer model.",[189,914,915,917],{},[193,916,412],{}," Identity, least privilege, fail-closed I\u002FO, and not turning the OS into a second store of the whole company. Security also cares that self-service configuration does not mean tenant-wide write keys.",[377,919,420],{"id":419},[189,921,922,925],{},[193,923,924],{},"“ChatGPT with integrations.”"," Plugins without scoped work, approval architecture, and durable decision records are plugins. A copilot with automation actions can move data. It cannot, by itself, make unapproved writes impossible.",[189,927,928,931,932,935,936,939],{},[193,929,930],{},"MLOps as a substitute."," MLOps governs ",[243,933,934],{},"model production",". An enterprise AI OS governs ",[243,937,938],{},"operational work that uses models",". They stack.",[189,941,942,945],{},[193,943,944],{},"Replacing the CRM."," Salesforce, NetSuite, Workday, and the warehouse remain authoritative. Duplicating them is a second system of record.",[189,947,948,951],{},[193,949,950],{},"OS as chatbot."," One user, one thread, no write path, no memory. That is a copilot with extra vocabulary.",[189,953,954,957],{},[193,955,956],{},"Forward-deployed as the only path."," Some warehouses need specialists. Most operators need to attach a connector and set a named signer in the UI.",[189,959,960],{},"Good looks like: workstreams, wiki, read-only-default connectors, agent teams, Lifecycle Graph, model routing, run quotes, fail-closed writes, self-service configuration. Failure looks like another model contract plus a six-month SOW.",[189,962,963,964,966,967,268],{},"For the copilot-versus-OS choice, see ",[207,965,586],{"href":585},". For vendor scoring, ",[207,968,970],{"href":969},"\u002Fblog\u002Fhow-to-evaluate-an-agent-harness\u002F","how to evaluate an agent harness",[216,972,470],{"id":469},[189,974,975],{},"Nimbus is a self-service enterprise AI OS. Operators configure it in the product.",[221,977,978,985,990,996,1001,1010],{},[224,979,980,984],{},[193,981,982],{},[207,983,31],{"href":788}," isolate process.",[224,986,987,989],{},[193,988,503],{}," holds asserted policy — approved playbooks, not a dump of PDFs a search might find.",[224,991,992,995],{},[193,993,994],{},"Connectors"," attach live systems. Default is read-only. Write-back is opt-in and gated.",[224,997,998,1000],{},[193,999,498],{}," are department-shaped.",[224,1002,1003,1005,1006,1009],{},[193,1004,23],{}," stores the causal record. ",[193,1007,1008],{},"Perception"," queries it in ordinary language.",[224,1011,1012,1015],{},[193,1013,1014],{},"Model routing"," puts routine extract on cheaper models.",[189,1017,490,1018,619,1021,1023,1024,499,1026,499,1028,499,1031,268],{},[207,1019,11],{"href":1020},"\u002Foverview\u002F",[207,1022,970],{"href":969},". Product surfaces: ",[207,1025,31],{"href":493},[207,1027,39],{"href":506},[207,1029,23],{"href":1030},"\u002Fproduct\u002Flifecycle-graph\u002F",[207,1032,1008],{"href":1033},"\u002Fproduct\u002Fperception\u002F",[216,1035,510],{"id":509},[377,1037,1039],{"id":1038},"is-an-enterprise-ai-os-just-chatgpt-with-integrations","Is an enterprise AI OS just “ChatGPT with integrations”?",[189,1041,1042],{},"No. Integrations without scoped work, approval architecture, and durable decision records are plugins. A copilot with automation actions can move data. It cannot, by itself, make unapproved writes impossible.",[377,1044,1046],{"id":1045},"how-is-this-different-from-mlops","How is this different from MLOps?",[189,1048,1049,1050,935,1052,1054],{},"MLOps governs ",[243,1051,934],{},[243,1053,938],{},". They stack. They do not substitute.",[377,1056,1058],{"id":1057},"do-we-still-need-a-crm-if-we-buy-an-os","Do we still need a CRM if we buy an OS?",[189,1060,1061],{},"Yes. Salesforce, NetSuite, Workday, and the warehouse remain authoritative.",[377,1063,1065],{"id":1064},"does-every-company-need-an-os","Does every company need an OS?",[189,1067,1068],{},"If the job is personal drafting with no writes to live systems, a governed copilot may be enough. The OS becomes the right abstraction when work crosses departments, when writes are material, and when you must reconstruct decisions.",[377,1070,1072],{"id":1071},"is-this-the-same-as-an-integration-platform-ipaas","Is this the same as an integration platform (iPaaS)?",[189,1074,1075],{},"No. iPaaS moves data on schedules and triggers. An AI OS runs language-using jobs with scope, spend, and a human gate. You may still need iPaaS. It does not quote a named signer on a CRM payload.",[377,1077,1079],{"id":1078},"does-operating-system-mean-we-install-software-on-laptops","Does “operating system” mean we install software on laptops?",[189,1081,1082],{},"No. It is a layer for work, not a desktop kernel. The metaphor is isolation, permissions, I\u002FO, and state.",[377,1084,1086],{"id":1085},"can-we-build-this-ourselves-on-a-model-api","Can we build this ourselves on a model API?",[189,1088,1089],{},"You can assemble pieces. You will still need isolation, connectors, gates, spend, and a graph. Most “we built a GPT” programmes stall at the copilot layer. McKinsey’s split between using AI and scaling it is that stall in survey form.",[377,1091,1093],{"id":1092},"where-do-agent-teams-fit","Where do agent teams fit?",[189,1095,1096,1097,268],{},"They are the department-shaped specialists the OS schedules onto workstreams. They are not the OS. See ",[207,1098,267],{"href":266},[377,1100,1102],{"id":1101},"how-does-nists-ai-rmf-map","How does NIST’s AI RMF map?",[189,1104,1105],{},"Govern (owners, policy), Map (inventory of jobs and systems), Measure (evidence, spend, rejects), Manage (fail-closed writes, incident path). A product can make those cheaper. A framework PDF cannot enforce them.",[377,1107,1109],{"id":1108},"what-is-perception-in-this-picture","What is Perception in this picture?",[189,1111,1112,1113,268],{},"Ordinary-language questions over the company’s graph, wiki, and scoped systems — with the next step being a workstream, not another search. See ",[207,1114,1008],{"href":1033},[377,1116,1118],{"id":1117},"do-we-need-a-forward-deployed-engineer-to-go-live","Do we need a forward-deployed engineer to go live?",[189,1120,1121],{},"Nothe default path. If operators cannot attach a read-only connector and set a named signer in the UI, you do not have a self-service OS. Specialists belong on genuine exceptions, such as a warehouse with no OAuth.",[377,1123,1125],{"id":1124},"is-search-rag-an-os","Is search (RAG) an OS?",[189,1127,1128,1129,619,1133,268],{},"No. Lookup-then-answer is infrastructure. It does not isolate jobs or gate writes. See ",[207,1130,1132],{"href":1131},"\u002Fblog\u002Fwhat-is-enterprise-rag\u002F","What is enterprise RAG",[207,1134,1136],{"href":1135},"\u002Fblog\u002Fnimbus-vs-glean\u002F","Nimbus vs Glean",[216,1138,613],{"id":612},[189,1140,1141,619,1143,268],{},[207,1142,789],{"href":788},[207,1144,902],{"href":901},[216,1146,626],{"id":625},[221,1148,1149,1154],{},[224,1150,1151],{},[207,1152,1153],{"href":731},"McKinsey, The state of AI (2025)",[224,1155,1156],{},[207,1157,1158],{"href":777},"NIST AI Risk Management Framework",{"title":169,"searchDepth":170,"depth":170,"links":1160},[1161,1162,1166,1167,1181,1182],{"id":218,"depth":170,"text":219},{"id":315,"depth":170,"text":316,"children":1163},[1164,1165],{"id":379,"depth":645,"text":380},{"id":419,"depth":645,"text":420},{"id":469,"depth":170,"text":470},{"id":509,"depth":170,"text":510,"children":1168},[1169,1170,1171,1172,1173,1174,1175,1176,1177,1178,1179,1180],{"id":1038,"depth":645,"text":1039},{"id":1045,"depth":645,"text":1046},{"id":1057,"depth":645,"text":1058},{"id":1064,"depth":645,"text":1065},{"id":1071,"depth":645,"text":1072},{"id":1078,"depth":645,"text":1079},{"id":1085,"depth":645,"text":1086},{"id":1092,"depth":645,"text":1093},{"id":1101,"depth":645,"text":1102},{"id":1108,"depth":645,"text":1109},{"id":1117,"depth":645,"text":1118},{"id":1124,"depth":645,"text":1125},{"id":612,"depth":170,"text":613},{"id":625,"depth":170,"text":626},{"eyebrow":667,"title":1184},"The layer between the model and the work",[1186,1189,1192],{"question":1187,"answer":1188},"What is an enterprise AI operating system?","The layer between the AI model and how departments actually work, the way an operating system sits between the chip and the applications.",{"question":1190,"answer":1191},"Is the model the operating system?","No. The model is a component. The operating system is how work, permissions, and records are organised around it.",{"question":1193,"answer":1194},"What should departments find there?","The same job, the same rules, and a record, instead of a different chatbot in every function.","\u002Fblog\u002Fwhat-is-an-enterprise-ai-operating-system",{"title":712,"description":720},"blog\u002Fwhat-is-an-enterprise-ai-operating-system",[681,685,1199,686],"operating-system","viSXGpdmqdd2fuaUmfH6OngdOgRbSEtUnbseaPgQvis",{"id":1202,"title":1203,"archived":163,"authors":1204,"badge":1206,"body":1207,"date":664,"definedTerm":164,"department":164,"description":1736,"extension":172,"eyebrow":164,"faqHeader":1737,"faqs":1739,"footerBand":164,"headline":164,"image":164,"industry":164,"jobType":164,"listed":131,"location":164,"navigation":131,"openRoles":164,"pageLayout":164,"path":1749,"relatedHeading":164,"seo":1750,"series":681,"sitemap":131,"status":164,"stem":1751,"subhead":164,"tags":1752,"video":164,"whyJoin":164,"workplaceType":164,"__hash__":1755},"content\u002Fblog\u002Fwhat-is-ai-governance.md","What is AI Governance",[1205],{"name":182,"to":183},{"label":185},{"type":166,"value":1208,"toc":1712},[1209,1216,1219,1222,1237,1240,1250,1252,1342,1352,1354,1362,1365,1396,1399,1427,1433,1435,1440,1450,1455,1460,1470,1472,1478,1484,1490,1498,1504,1511,1528,1530,1533,1540,1553,1560,1562,1566,1574,1578,1581,1585,1592,1596,1599,1603,1606,1610,1617,1621,1624,1628,1631,1635,1638,1642,1645,1649,1654,1658,1661,1663,1672,1674],[189,1210,1211,1212,1215],{},"AI governance is the set of rules, ",[193,1213,1214],{},"enforced in the software people actually use",", that decide who may use which AI, on which company data, and whether that AI is allowed to change a live business system — plus a record of what happened afterwards.",[189,1217,1218],{},"A training video is not that. An acceptable-use PDF is not that. An admin toggle the model can ignore is not that. If an unapproved change can still succeed, you have guidance, not governance.",[189,1220,1221],{},"People use the phrase for three different things, and they get mixed up:",[812,1223,1224,1231,1234],{},[224,1225,1226,1227,268],{},"A public commitment — for example the ",[207,1228,1230],{"href":1229},"https:\u002F\u002Foecd.ai\u002Fen\u002Fai-principles","OECD AI Principles",[224,1232,1233],{},"A company committee with a risk register.",[224,1235,1236],{},"The runtime that actually stops a change to CRM, ERP, or a customer message.",[189,1238,1239],{},"All three are real. Only the third one would have blocked an unlogged field change that later showed up in a forecast.",[189,1241,1242,1246,1247,1249],{},[207,1243,1245],{"href":1244},"https:\u002F\u002Fwww.gartner.com\u002Fen\u002Farticles\u002Fai-governance-trism","Gartner’s TRiSM"," language is about that third layer: trust, risk, and security around the systems that run — not a quarterly slide about principles. The ",[207,1248,1158],{"href":777}," says the same thing in public-sector language: Govern, Map, Measure, Manage. Mapping systems and measuring incidents still fail if the product people click can write to Salesforce without a named signer.",[216,1251,219],{"id":218},[221,1253,1254,1260,1273,1284,1289,1295,1310,1319,1330],{},[224,1255,1256,1259],{},[193,1257,1258],{},"Live business system."," CRM, ERP, HR, billing — the tools that hold official numbers and customer records. At work, this is where a fluent sentence becomes a fact other teams will inherit.",[224,1261,1262,1265,1266,1269,1270,1272],{},[193,1263,1264],{},"Write \u002F write-back."," The AI is allowed to ",[243,1267,1268],{},"change"," that system, not only draft a suggestion. See ",[207,1271,799],{"href":798},". At work, a next-step note and an Amount field are not the same risk class.",[224,1274,1275,1278,1279,1283],{},[193,1276,1277],{},"Human-in-the-loop."," A person must approve before the job can finish. See ",[207,1280,1282],{"href":1281},"\u002Fblog\u002Fwhat-is-human-in-the-loop-ai\u002F","What is human-in-the-loop AI",". At work, the gate shows the payload in the language of the live system, not a wall of prompt text.",[224,1285,1286,1288],{},[193,1287,273],{}," The identity that authorised the change. At work, “someone in the channel clicked yes” is not a signer.",[224,1290,1291,1294],{},[193,1292,1293],{},"Fail-closed."," Missing approval means nothing happens. Fail-open means the change goes through unless someone happens to stop it.",[224,1296,1297,1300,1301,1304,1305,1309],{},[193,1298,1299],{},"DPIA."," A data-protection impact assessment — thinking through purpose, risk, and personal data ",[243,1302,1303],{},"before"," you turn a tool loose. ",[207,1306,1308],{"href":1307},"https:\u002F\u002Fico.org.uk\u002Ffor-organisations\u002Fuk-gdpr-guidance-and-resources\u002Fartificial-intelligence\u002F","UK ICO guidance on AI and data protection"," still wants a lawful basis and purpose when the “user” is an AI.",[224,1311,1312,1315,1316,268],{},[193,1313,1314],{},"Shadow AI."," Personal ChatGPT for work because the official path is missing. See ",[207,1317,1318],{"href":567},"What is shadow AI",[224,1320,1321,1324,1325,1329],{},[193,1322,1323],{},"Inventory."," A list of where AI actually runs. The ",[207,1326,1328],{"href":1327},"https:\u002F\u002Fwww.justice.gov\u002Fmedia\u002F1373026\u002Fdl","US plan described in OMB M-24-10"," puts a named owner and an inventory first, not a PDF.",[224,1331,1332,241,1335,1337,1338,1341],{},[193,1333,1334],{},"Least privilege.",[243,1336,245],{}," job. A ",[207,1339,1340],{"href":788},"workstream"," is how that instinct becomes a company object.",[189,1343,1344,1345,1348,1349,1351],{},"Model safety is adjacent and different. Safety is about what a model will say in the abstract. Enterprise governance is about what ",[243,1346,1347],{},"your"," people and tools may do with ",[243,1350,1347],{}," systems and data. You can have a carefully aligned model and still have ungoverned CRM writes.",[216,1353,316],{"id":315},[189,1355,1356,1357,1361],{},"Without working rules, AI becomes a side effect. A field moves. A journal posts. A customer is told a policy the company does not hold. Nobody can say who allowed it. In February 2024, a British Columbia tribunal held Air Canada responsible for a chatbot that invented a bereavement-fare policy. ",[207,1358,1360],{"href":1359},"https:\u002F\u002Fwww.cbc.ca\u002Fnews\u002Fcanada\u002Fbritish-columbia\u002Fair-canada-chatbot-lawsuit-1.7116416","CBC reported"," that the airline’s argument — the chatbot is a separate legal entity — failed. A customer-facing commitment without a working gate is still the company’s commitment.",[189,1363,1364],{},"It affects you if you:",[221,1366,1367,1373,1379,1390],{},[224,1368,1369,1372],{},[193,1370,1371],{},"Own a number."," Forecasts and close packs inherit whatever changed.",[224,1374,1375,1378],{},[193,1376,1377],{},"Own a customer relationship."," Model output that becomes a commitment is still the company’s commitment.",[224,1380,1381,1384,1385,1389],{},[193,1382,1383],{},"Own risk or legal."," Privacy law does not pause for a chatbot. ",[207,1386,1388],{"href":1387},"https:\u002F\u002Feur-lex.europa.eu\u002Feli\u002Freg\u002F2016\u002F679\u002Foj","GDPR"," still applies to purpose, minimisation, and erasure.",[224,1391,1392,1395],{},[193,1393,1394],{},"Are asked “who is in charge of AI here?”"," An inventory and a named owner beat a principles slide.",[189,1397,1398],{},"Good governance in practice is four working rules:",[221,1400,1401,1407,1413,1419],{},[224,1402,1403,1406],{},[193,1404,1405],{},"People and rights."," Humans and AI tools are both actors. Roles decide what they may start, see, and sign.",[224,1408,1409,1412],{},[193,1410,1411],{},"Data at question time."," Purpose and minimisation still apply when an AI is the one looking.",[224,1414,1415,1418],{},[193,1416,1417],{},"Action rights."," Read-only is a control. Unrestricted tools are an incident waiting for a bad prompt.",[224,1420,1421,1424,1425,268],{},[193,1422,1423],{},"Evidence and spend."," Chat scrollback is not a management system. Uncapped spend is a budget failure and often a security failure. See ",[207,1426,284],{"href":283},[189,1428,1429,1430,1432],{},"Blocking consumer ChatGPT at the office network, while people use personal phones, is not governance. It is a ",[207,1431,568],{"href":567}," problem with extra steps.",[377,1434,380],{"id":379},[189,1436,1437,1439],{},[193,1438,385],{}," Governance is whether an AI-proposed journal can post, against which checklist, with which signer, and whether the spend of the run was capped. “Unlimited AI” is not a control. Surprise inference bills are a governance failure that looks like a cloud invoice.",[189,1441,1442,1444,1445,1449],{},[193,1443,391],{}," Lawful basis, purpose limitation, customer-facing language, and reconstructable authorisation. Legal also has to separate the OECD-style public commitment from the runtime. A principles page does not implement Article-style oversight. For higher-risk systems, ",[207,1446,1448],{"href":1447},"https:\u002F\u002Feur-lex.europa.eu\u002Feli\u002Freg\u002F2024\u002F1689\u002Foj","EU AI law"," Article 14 talks about effective oversight: people must be able to interpret outputs and interrupt the system. A footer that says “generated by AI” is not that.",[189,1451,1452,1454],{},[193,1453,400],{}," Isolation of jobs, connector scope, and a place to put a paused run. Ops already runs change control. Governance is change control that includes a model as a proposer.",[189,1456,1457,1459],{},[193,1458,406],{}," The difference between a draft email and a sent commitment; between a suggested next step and a changed Amount. GTM feels friction first. The honest metric is time-to-approved-write, not time-to-first-answer.",[189,1461,1462,1464,1465,1469],{},[193,1463,412],{}," Identity of the connected user, read versus write, prompt injection as a path to a tool call, and the new store created by logs and indexes. The ",[207,1466,1468],{"href":1467},"https:\u002F\u002Fgenai.owasp.org\u002Fllm-top-10\u002F","OWASP Top 10 for LLM applications"," treats retrieval and tool use as a security surface, not only a quality issue. Network DLP helps with paste-out. It does not quote a CRM change.",[377,1471,420],{"id":419},[189,1473,1474,1477],{},[193,1475,1476],{},"Governance as a committee."," Useful for risk registers. Useless if the product can still write.",[189,1479,1480,1483],{},[193,1481,1482],{},"Governance as model safety."," Refusals on public-web questions do not bind Salesforce.",[189,1485,1486,1489],{},[193,1487,1488],{},"Governance as a secure web gateway."," Necessary for some paste-out paths. Insufficient for writes, approvals, and causal history.",[189,1491,1492,1495,1496,268],{},[193,1493,1494],{},"Governance as blocking."," Blocks without a sanctioned path train people onto phones. See ",[207,1497,1318],{"href":567},[189,1499,1500,1503],{},[193,1501,1502],{},"Theatre."," A checkbox, a prompt that says “ask first,” or an admin toggle the model can ignore.",[189,1505,1506,1507,1510],{},"Good looks like: connectors default to read-only; writes are quoted; a named signer cannot be waived by the model; evidence lives on a ",[207,1508,1509],{"href":303},"lifecycle graph","; spend has a ceiling; scope follows the job. Failure looks like a PDF, a blocked URL, and a personal API key in a wiki.",[189,1512,1513,1514,1517,1518,1521,1522,1524,1525,1527],{},"Adjacent concepts: ",[207,1515,1516],{"href":798},"write-back governance"," is the write subset. ",[207,1519,1520],{"href":1281},"Human-in-the-loop"," is the gate. ",[207,1523,31],{"href":788}," are the isolation unit. An ",[207,1526,486],{"href":464}," is the product shape that makes those rules the default path.",[216,1529,470],{"id":469},[189,1531,1532],{},"Nimbus treats governance as how work is released, not as a sidecar policy engine.",[189,1534,1535,1536,1539],{},"Connectors — secure links to live systems — default to ",[193,1537,1538],{},"read-only",". When a change is proposed, the product shows the intended action and waits. A named person must sign. The model cannot waive the gate. Missing approval is fail-closed: nothing happens.",[189,1541,1542,1543,1545,1546,1548,1549,1552],{},"Scope is the ",[207,1544,1340],{"href":788},": one job, with the playbooks, systems, teams, and budget that belong to that job. Evidence is the ",[207,1547,23],{"href":303},". The ",[207,1550,1551],{"href":293},"company wiki"," is the asserted policy the run must cite. Model routing does not bypass the gate.",[189,1554,490,1555,1557,1558,268],{},[207,1556,39],{"href":506},". For scoring vendors: ",[207,1559,970],{"href":969},[216,1561,510],{"id":509},[377,1563,1565],{"id":1564},"is-ai-governance-the-same-as-making-the-model-safe","Is AI governance the same as making the model “safe”?",[189,1567,1568,1569,1348,1571,1573],{},"No. Model safety is about what the model will say in the abstract. Enterprise governance is about what ",[243,1570,1347],{},[243,1572,1347],{}," systems and data.",[377,1575,1577],{"id":1576},"can-we-rely-on-the-secure-web-gateway","Can we rely on the secure web gateway?",[189,1579,1580],{},"Network controls help with paste-out. They do not quote a CRM change, bind an approver, or store a causal history. Use both.",[377,1582,1584],{"id":1583},"must-a-person-always-approve","Must a person always approve?",[189,1586,1587,1588,1591],{},"For many operational writes, yes. For read-only analysis, maybe not. The mistake is calling a system “human-approved” because a human ",[243,1589,1590],{},"could"," look, while changes proceed on model initiative.",[377,1593,1595],{"id":1594},"do-the-oecd-ai-principles-require-a-specific-product","Do the OECD AI Principles require a specific product?",[189,1597,1598],{},"No. They are a public commitment. A product can make evidence cheaper to produce. The commitment does not implement a gate.",[377,1600,1602],{"id":1601},"is-a-dpia-enough-to-go-live","Is a DPIA enough to go live?",[189,1604,1605],{},"It is necessary thinking, not a runtime. You still need identity, scope, fail-closed writes, and a record. The DPIA should describe those controls, not replace them.",[377,1607,1609],{"id":1608},"does-blocking-chatgpt-count-as-governance","Does blocking ChatGPT count as governance?",[189,1611,1612,1613,1616],{},"It is a network control. Without a sanctioned path that can see the right files, people use personal phones. Blocking can tighten ",[243,1614,1615],{},"after"," substitution exists.",[377,1618,1620],{"id":1619},"how-is-this-different-from-it-change-management","How is this different from IT change management?",[189,1622,1623],{},"It is the same instinct — who may change production, with what evidence — applied to a proposer that speaks English. Existing CAB processes rarely see model-initiated payloads unless the product emits them.",[377,1625,1627],{"id":1626},"who-should-be-the-named-owner-of-ai","Who should be the named owner of AI?",[189,1629,1630],{},"Someone who can inventory systems and stop a write path, not a volunteer “champion” with no authority over CRM. Federal-style guidance starts with inventory and ownership for a reason.",[377,1632,1634],{"id":1633},"can-we-govern-only-customer-facing-chatbots-and-ignore-internal-copilots","Can we govern only customer-facing chatbots and ignore internal copilots?",[189,1636,1637],{},"Internal tools still process personal data and still write to live systems. Air Canada was customer-facing. Ungoverned CRM hygiene is an internal path to the same class of invented fact.",[377,1639,1641],{"id":1640},"do-we-need-the-eu-ai-act-if-we-are-not-a-high-risk-provider","Do we need the EU AI Act if we are not a high-risk provider?",[189,1643,1644],{},"You may still have GDPR duties, sector rules, and customer contracts. Oversight and records are useful even when a specific Act title does not apply. Do not claim “Act compliant” because you have a button.",[377,1646,1648],{"id":1647},"where-does-spend-fit","Where does spend fit?",[189,1650,1651,1652,268],{},"Uncapped inference is a control failure. Quotes, ceilings, and attribution by job are governance of a scarce, abusable resource. See ",[207,1653,284],{"href":283},[377,1655,1657],{"id":1656},"is-an-acceptable-use-policy-still-worth-writing","Is an acceptable-use policy still worth writing?",[189,1659,1660],{},"Yes, as communication. No, as enforcement. Write the PDF. Then put the same rules in the product people actually use.",[216,1662,613],{"id":612},[189,1664,1665,499,1667,1669,1670,268],{},[207,1666,799],{"href":798},[207,1668,1318],{"href":567},", and ",[207,1671,618],{"href":464},[216,1673,626],{"id":625},[221,1675,1676,1681,1685,1690,1694,1699,1703,1708],{},[224,1677,1678],{},[207,1679,1680],{"href":1244},"Gartner, AI governance and TRiSM",[224,1682,1683],{},[207,1684,1230],{"href":1229},[224,1686,1687],{},[207,1688,1689],{"href":1307},"ICO, AI and data protection",[224,1691,1692],{},[207,1693,1158],{"href":777},[224,1695,1696],{},[207,1697,1698],{"href":1359},"CBC, Air Canada chatbot lawsuit",[224,1700,1701],{},[207,1702,1388],{"href":1387},[224,1704,1705],{},[207,1706,1707],{"href":1447},"EU AI Act (Regulation 2024\u002F1689)",[224,1709,1710],{},[207,1711,1468],{"href":1467},{"title":169,"searchDepth":170,"depth":170,"links":1713},[1714,1715,1719,1720,1734,1735],{"id":218,"depth":170,"text":219},{"id":315,"depth":170,"text":316,"children":1716},[1717,1718],{"id":379,"depth":645,"text":380},{"id":419,"depth":645,"text":420},{"id":469,"depth":170,"text":470},{"id":509,"depth":170,"text":510,"children":1721},[1722,1723,1724,1725,1726,1727,1728,1729,1730,1731,1732,1733],{"id":1564,"depth":645,"text":1565},{"id":1576,"depth":645,"text":1577},{"id":1583,"depth":645,"text":1584},{"id":1594,"depth":645,"text":1595},{"id":1601,"depth":645,"text":1602},{"id":1608,"depth":645,"text":1609},{"id":1619,"depth":645,"text":1620},{"id":1626,"depth":645,"text":1627},{"id":1633,"depth":645,"text":1634},{"id":1640,"depth":645,"text":1641},{"id":1647,"depth":645,"text":1648},{"id":1656,"depth":645,"text":1657},{"id":612,"depth":170,"text":613},{"id":625,"depth":170,"text":626},"AI governance is the working rules for who may use which AI, on which data, and whether it may change a live business system.",{"eyebrow":667,"title":1738},"Rules that actually stop a change",[1740,1743,1746],{"question":1741,"answer":1742},"What is AI governance?","The working rules for who may use which AI, on which data, and whether it may change a live business system, plus a record of what happened.",{"question":1744,"answer":1745},"Is a policy PDF governance?","Only if the software enforces it. A document the model can ignore is guidance.",{"question":1747,"answer":1748},"What has to be true before a write?","A named person or a named rule has allowed that exact change. If they have not, the change does not happen.","\u002Fblog\u002Fwhat-is-ai-governance",{"title":1203,"description":1736},"blog\u002Fwhat-is-ai-governance",[681,686,1753,1754],"compliance","audit","3P7kWkstCX-Me3qq2PZcPS9sMhi7EacGegrGNsiBExQ",{"enabled":163,"message":1757,"linkLabel":79,"linkHref":80,"id":1758,"title":1759,"archived":163,"authors":164,"badge":164,"body":1760,"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":1764,"relatedHeading":164,"seo":1765,"series":164,"sitemap":163,"status":164,"stem":1766,"subhead":164,"tags":164,"video":164,"whyJoin":164,"workplaceType":164,"__hash__":1767},"We're hiring! Join the team building the Sentient Enterprise.","content\u002Fshared\u002Fhiring.md","Hiring banner",{"type":166,"value":1761,"toc":1762},[],{"title":169,"searchDepth":170,"depth":170,"links":1763},[],"\u002Fshared\u002Fhiring",{"title":1759,"description":169},"shared\u002Fhiring","1zs3boivKda1e-b-hAyuNcmZSKjZUAXmecnwHVgcHzk",{"fold":1769,"id":1773,"title":1774,"archived":163,"authors":164,"badge":164,"body":1775,"date":164,"definedTerm":164,"department":164,"description":169,"extension":172,"eyebrow":164,"faqHeader":164,"faqs":164,"footerBand":1779,"headline":164,"image":164,"industry":164,"jobType":164,"listed":131,"location":164,"navigation":131,"openRoles":164,"pageLayout":164,"path":1783,"relatedHeading":164,"seo":1784,"series":164,"sitemap":163,"status":164,"stem":1785,"subhead":164,"tags":164,"video":164,"whyJoin":164,"workplaceType":164,"__hash__":1786},{"headline":1770,"description":1771,"primaryLabel":8,"primaryTo":1772,"secondaryLabel":704,"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":1776,"toc":1777},[],{"title":169,"searchDepth":170,"depth":170,"links":1778},[],{"headline":1780,"description":1781,"primaryLabel":8,"primaryTo":1772,"secondaryLabel":1782,"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":1774,"description":169},"shared\u002Fcta","eYqahyaPnbp8GKrWpoORbZdtmkWmgHr5F61ZHOnb8sY",1792999485768]