[{"data":1,"prerenderedAt":1745},["ShallowReactive",2],{"site-nav-content":3,"blog:/blog/what-is-a-lifecycle-graph":163,"blog-index-copy":672,"blog:/blog/what-is-a-lifecycle-graph:surround":693,"hiring-banner-content":1714,"site-cta-content":1726},{"header":4,"productNav":9,"nav":42,"footer":61,"askAI":116,"id":147,"title":148,"archived":149,"authors":150,"badge":150,"body":151,"date":150,"department":150,"description":155,"extension":158,"eyebrow":150,"faqHeader":150,"faqs":150,"footerBand":150,"headline":150,"image":150,"industry":150,"jobType":150,"listed":149,"location":150,"navigation":115,"openRoles":150,"pageLayout":150,"path":159,"relatedHeading":150,"seo":160,"series":150,"sitemap":115,"status":150,"stem":161,"subhead":150,"tags":150,"video":150,"whyJoin":150,"workplaceType":150,"__hash__":162},{"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","/overview","Seven layers. One closed loop.",{"label":15,"to":16,"description":17},"Conflux","/product/conflux","Where your team, workstreams, and agents meet.",{"label":19,"to":20,"description":21},"Agent Teams","/product/agent-teams","Specialist teams - governed from day one.",{"label":23,"to":24,"description":25},"Lifecycle Graph","/product/lifecycle-graph","Intelligence that compounds across every interaction.",{"label":27,"to":28,"description":29},"Company Wiki","/product/wiki","Playbooks and policies where expertise stays.",{"label":31,"to":32,"description":33},"Workstreams","/product/workstreams","From brief to signed-off deliverable on one canvas.",{"label":35,"to":36,"description":37},"Perception Console","/product/perception","Ask your whole business in plain English.",{"label":39,"to":40,"description":41},"Governance","/product/governance","Frontier AI you can actually sign off on.",[43,46,49,52,55,58],{"label":44,"to":45},"Models","/models",{"label":47,"to":48},"Pricing","/pricing",{"label":50,"to":51},"Integrations","/integrations",{"label":53,"to":54},"Security","/security",{"label":56,"to":57},"Partners","/partners",{"label":59,"to":60},"Insights","/blog",{"productHeading":5,"companyHeading":62,"legalHeading":63,"docsLabel":64,"docsUrl":65,"statementLines":66,"copyright":69,"companyLinks":70,"legalLinks":85,"socialLinks":95,"bottomLinks":105},"Company","Legal","Docs","https://docs.gonimbus.ai",[67,68],"Stop training someone else's model.","Control your AI.","© 2026 Nimbus Intelligence, Inc. All rights reserved.",[71,72,73,74,75,77,80,83],{"label":47,"to":48},{"label":50,"to":51},{"label":53,"to":54},{"label":59,"to":60},{"label":76,"to":57},"Partner Program",{"label":78,"to":79},"Careers","/careers",{"label":81,"to":82},"System status","/status",{"label":7,"to":84},"/contact",[86,89,92],{"label":87,"to":88},"Terms of Service","/terms",{"label":90,"to":91},"Privacy Policy","/privacy",{"label":93,"to":94},"Compliance","/compliance",[96,99,102],{"label":97,"href":98},"LinkedIn","https://www.linkedin.com/company/gonimbusai/",{"label":100,"href":101},"X","https://x.com/gonimbusai",{"label":103,"href":104},"Instagram","https://www.instagram.com/gonimbus_ai/",[106,108,110,111,112],{"label":107,"to":88},"Terms",{"label":109,"to":91},"Privacy",{"label":93,"to":94},{"label":81,"to":82},{"label":113,"to":114,"external":115},"LLMs.txt","/llms.txt",true,{"text":117,"prompt":118},"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":119,"platforms":121},{" Summarize the highlights from Nimbus's website":120},"https://gonimbus.ai",[122,127,132,137,142],{"name":123,"label":124,"icon":125,"hrefPrefix":126},"chatgpt","ChatGPT","simple-icons:openai","https://chatgpt.com/?prompt=",{"name":128,"label":129,"icon":130,"hrefPrefix":131},"perplexity","Perplexity","mdi:magnify","https://www.perplexity.ai/search/new?q=",{"name":133,"label":134,"icon":135,"hrefPrefix":136},"grok","Grok","simple-icons:x","https://x.com/i/grok?text=",{"name":138,"label":139,"icon":140,"hrefPrefix":141},"claude","Claude","simple-icons:anthropic","https://claude.ai/new?q=",{"name":143,"label":144,"icon":145,"hrefPrefix":146},"google-ai","Google AI","simple-icons:google","https://www.google.com/search?udm=50&aep=11&q=","content/shared/nav.md","Site navigation",false,null,{"type":152,"value":153,"toc":154},"minimark",[],{"title":155,"searchDepth":156,"depth":156,"links":157},"",2,[],"md","/shared/nav",{"title":148,"description":155},"shared/nav","p6IsjEfjcrwkjwvGsPcuEkpoJSeUn2vJXPTxLWQfw9M",{"id":164,"title":165,"archived":149,"authors":166,"badge":169,"body":171,"date":661,"department":150,"description":662,"extension":158,"eyebrow":150,"faqHeader":150,"faqs":150,"footerBand":150,"headline":150,"image":150,"industry":150,"jobType":150,"listed":149,"location":150,"navigation":115,"openRoles":150,"pageLayout":150,"path":663,"relatedHeading":150,"seo":664,"series":665,"sitemap":115,"status":150,"stem":666,"subhead":150,"tags":667,"video":150,"whyJoin":150,"workplaceType":150,"__hash__":671},"content/blog/what-is-a-lifecycle-graph.md","What is a Lifecycle Graph",[167],{"name":168,"to":120},"Nimbus Research",{"label":170},"Explainer",{"type":152,"value":172,"toc":635},[173,182,185,188,191,196,256,278,282,285,288,291,317,332,337,348,354,360,366,372,376,379,396,399,402,411,415,422,433,440,449,452,456,462,468,475,488,492,496,503,507,513,517,520,524,527,531,538,542,545,549,552,556,562,566,569,573,576,580,585,589,596,600,610,614],[174,175,176,177,181],"p",{},"When people ask what a lifecycle graph is, they are usually asking about ",[178,179,180],"strong",{},"causality",": why did this number, field, or decision change?",[174,183,184],{},"Causality is the difference between “two things happened around the same time” and “this caused that.” If pipeline coverage went up in the same month AI usage went up, that is a coincidence until you can show the actual steps: what was asked, what was used, who approved it, and what the live system did.",[174,186,187],{},"A lifecycle graph is the company’s record of those steps. It is not a chat history. Chat history shows that someone talked to a model. A graph shows the chain from the question to the outcome, so the next person — or an auditor — can follow it.",[174,189,190],{},"This is an operations problem that existed before generative AI. ERP journals already needed authorisation trails. CRM already had field history. What changed is that a new kind of actor can now propose, and sometimes execute, those changes in fluent language. If the “why” lives only in a personal chat, the company has a causality gap the moment that person leaves, the vendor rotates logs, or the model version rolls.",[192,193,195],"h2",{"id":194},"words-youll-hear","Words you’ll hear",[197,198,199,206,212,226,232,238,244,250],"ul",{},[200,201,202,205],"li",{},[178,203,204],{},"Causality."," Being able to say what caused what, with evidence. At work, this is “this next-step field changed because this brief ran, cited this playbook version, and this person signed this payload.”",[200,207,208,211],{},[178,209,210],{},"Correlation."," Two things moving together. Not the same as cause. At work, AI usage and pipeline moving in the same quarter is a slide, not an explanation.",[200,213,214,217,218,225],{},[178,215,216],{},"Provenance."," The trail of who, what, and when behind a piece of data. ",[219,220,224],"a",{"href":221,"rel":222},"https://www.w3.org/TR/prov-overview/",[223],"nofollow","W3C PROV"," is the open standard for that idea: entities, activities, and agents, linked so you can reconstruct derivation. A lifecycle graph is that instinct applied to AI-mediated work, not a claim that you have implemented the full W3C stack.",[200,227,228,231],{},[178,229,230],{},"System of record."," The official live system that holds the fact — Salesforce for an opportunity, NetSuite for a journal. The graph should point at that record, not become a second copy of it. At work, pointing is how you avoid a second CRM that nobody can delete.",[200,233,234,237],{},[178,235,236],{},"Audit trail."," A log that something happened. Useful, but thin if it cannot join the question, the policy, the signer, and the change. Provider API logs are an audit trail of calls. They are not a story of the job.",[200,239,240,243],{},[178,241,242],{},"Lineage."," Which sources fed which proposal. At work, “which wiki version and which CRM records were in scope when this quote was generated?”",[200,245,246,249],{},[178,247,248],{},"Retention."," How long a class of node is kept. A journal that feeds the books may need years. A draft may need weeks.",[200,251,252,255],{},[178,253,254],{},"Scope."," Which job’s chain you are allowed to see. At work, a go-to-market question should not surface People Ops briefs.",[174,257,258,259,262,263,267,268,272,273,277],{},"Keep this graph apart from two neighbours. A ",[178,260,261],{},"business knowledge graph"," models customers, products, and sites. ",[219,264,266],{"href":265},"what-is-enterprise-rag","Enterprise RAG"," retrieves documents that ",[269,270,271],"em",{},"exist",". ",[219,274,276],{"href":275},"what-is-institutional-memory-in-enterprise-ai","Institutional memory"," is the broader goal — what the company still knows after people leave. The lifecycle graph is the decision-memory layer of that goal: the chain of AI-mediated work.",[192,279,281],{"id":280},"why-you-should-care","Why you should care",[174,283,284],{},"AI makes it easy to change company systems without leaving a story. Someone asks a model to tidy CRM notes. A field moves. Next quarter, finance or legal asks why. The person who asked has left. The “why” lived in a personal chat. The CRM only shows the new value.",[174,286,287],{},"That is a causality problem. You cannot manage what you cannot reconstruct.",[174,289,290],{},"It affects you if you:",[197,292,293,299,305,311],{},[200,294,295,298],{},[178,296,297],{},"Sign off on numbers."," Forecasts, journals, and board packs inherit whatever AI changed last month.",[200,300,301,304],{},[178,302,303],{},"Inherit someone else’s work."," You need the exception, not a rumour that “we always do 18% for strategic accounts.”",[200,306,307,310],{},[178,308,309],{},"Answer auditors or regulators."," They will not accept “the chatbot did it.”",[200,312,313,316],{},[178,314,315],{},"Switch vendors or models."," Provider logs are the vendor’s artefact. They are not your company memory.",[174,318,319,320,323,324,328,329],{},"This is not the same as proving that a discount ",[269,321,322],{},"caused"," a won deal. That is a statistics question. See ",[219,325,327],{"href":326},"what-is-causal-ai-for-operations","What is causal AI for operations",". A lifecycle graph answers a more basic one: ",[178,330,331],{},"what did we actually do, and who caused it?",[333,334,336],"h3",{"id":335},"what-changes-by-role","What changes by role",[174,338,339,342,343,347],{},[178,340,341],{},"Finance."," Close packs and forecasts inherit field history. If an AI-proposed journal posted, finance needs the brief, the playbook version, the named signer, and the ERP response — not a Slack screenshot. Spend also belongs on the chain: a run that stopped because a cap was hit is a causal fact, not a missing invoice. See ",[219,344,346],{"href":345},"what-is-ai-token-economics","What is AI token economics",".",[174,349,350,353],{},[178,351,352],{},"Legal."," Exception language, customer commitments, and “who saw what” are discovery questions. A graph that points at the payload the signer saw is evidence. A chat export from a personal account is not. Legal also cares about retention and deletion: infinite chat fails a privacy review; typed retention with export and legal hold is how records programmes already work.",[174,355,356,359],{},[178,357,358],{},"Operations."," Handoffs fail when the next shift cannot see why a run paused. Human wait is a node, not an interruption. Incident reviews need the same chain: which connector was read-only, which write was refused, which wiki page caused the flag.",[174,361,362,365],{},[178,363,364],{},"Go-to-market."," Pipeline hygiene and renewal exceptions are where “the bot updated it” becomes a forecast problem. GTM needs to see the quoted fields, not a summary that says “updated pricing.” They also need scope: one team’s competitive notes should not leak into another region’s chain.",[174,367,368,371],{},[178,369,370],{},"Security."," The graph is a sensitive store. It should not hold full transcripts with secrets by default, other teams’ out-of-scope work, or the model’s private scratch reasoning. Access control on the graph is as important as access control on the CRM. A query surface that ignores vaults recreates the god workspace.",[333,373,375],{"id":374},"what-belongs-on-the-chain","What belongs on the chain",[174,377,378],{},"Keep the links that let a non-engineer reconstruct a change:",[197,380,381,384,387,390,393],{},[200,382,383],{},"the job and the question",[200,385,386],{},"the sources (which playbook version, which records, which files)",[200,388,389],{},"the people (the model is not an answer for “who”)",[200,391,392],{},"the proposed change, in the language of the live system — fields and values, not “updated pricing”",[200,394,395],{},"the decision, the timestamp, and whether the live system accepted it",[174,397,398],{},"Point at the CRM record and the policy page. Do not copy the whole company into the graph. Copies become a second official system, and a deletion problem.",[174,400,401],{},"Do not keep, by default: the model’s private scratch reasoning, full transcripts with secrets, or other teams’ work that was never in scope.",[174,403,404,405,410],{},"Retention should follow the type of record. A journal that feeds the books may need years. A draft may need weeks. “Keep everything forever because AI” fails a privacy review. ",[219,406,409],{"href":407,"rel":408},"https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/artificial-intelligence/",[223],"UK ICO guidance on AI and data protection"," still wants purpose and minimisation when the “user” is an AI.",[333,412,414],{"id":413},"what-people-get-wrong","What people get wrong",[174,416,417,418,421],{},"The first mistake is ",[178,419,420],{},"treating chat history as the record",". Chat is a user interface. It is not a join of brief, policy, signer, and system response.",[174,423,424,425,428,429,432],{},"The second is ",[178,426,427],{},"retrofitting",". Copying six months of ChatGPT and Slack into a warehouse is archaeology. You still need something that ",[269,430,431],{},"emits"," events at the moment of the brief, the quote, and the approval.",[174,434,435,436,439],{},"The third is ",[178,437,438],{},"a second CRM",". Duplicating every opportunity into the graph “for completeness” creates conflicting official numbers and an erasure nightmare.",[174,441,442,443,446,447,347],{},"The fourth is ",[178,444,445],{},"confusing this with causal science",". A fluent model paragraph that says “because” is not identification. Neither is a dashboard of two rising lines. See ",[219,448,327],{"href":326},[174,450,451],{},"Good looks like reconstructable interventions with pointers, named people, and typed retention. Failure looks like a vendor log, a personal thread, or an infinite lake of tokens.",[192,453,455],{"id":454},"how-this-shows-up-in-nimbus","How this shows up in Nimbus",[174,457,458,459,461],{},"Nimbus’s ",[178,460,23],{}," is that chain as a product: briefs, playbook citations, connector reads, spend, approvals, and write results are linked as work proceeds.",[174,463,464,467],{},[178,465,466],{},"Perception"," is how you ask it in ordinary language — “why did this opportunity change last month?” — instead of reconstructing Slack.",[174,469,470,471,474],{},"The graph records ",[178,472,473],{},"work",", not every token the company ever sent to a model. Scope follows the job, so a go-to-market question should not surface People Ops briefs. Connectors default to read-only; a read that did not write is itself a node worth knowing. Fail-closed writes mean a missing named signer is a recorded refusal, not a silent mutation.",[174,476,477,478,480,481,483,484,347],{},"Product: ",[219,479,23],{"href":24}," and ",[219,482,466],{"href":36},". The job that produces the chain is a ",[219,485,487],{"href":486},"what-is-an-ai-workstream","workstream",[192,489,491],{"id":490},"questions-people-actually-ask","Questions people actually ask",[333,493,495],{"id":494},"is-this-just-a-knowledge-graph-of-the-business","Is this just a knowledge graph of the business?",[174,497,498,499,502],{},"No. A business knowledge graph models customers, products, and sites. A lifecycle graph models ",[178,500,501],{},"AI-mediated work"," — what was asked, who signed, what changed. They can link (the write points at an opportunity). They are not the same thing.",[333,504,506],{"id":505},"cant-the-warehouse-be-the-record","Can’t the warehouse be the record?",[174,508,509,510,512],{},"You can copy events into a warehouse for reporting. You still need something that ",[269,511,431],{}," those events at the moment of the brief, the quote, and the approval. Retrofitting six months of ChatGPT and Slack is archaeology, not operations.",[333,514,516],{"id":515},"how-is-this-different-from-the-model-providers-logs","How is this different from the model provider’s logs?",[174,518,519],{},"Provider logs show API calls. They do not know your job, your playbook version, your approver, or whether the write was rejected.",[333,521,523],{"id":522},"how-long-should-we-keep-it","How long should we keep it?",[174,525,526],{},"Treat it like other control evidence. Align retention with the type of record, legal hold, and storage limits. The product must support export, deletion, and access control — not infinite chat.",[333,528,530],{"id":529},"how-does-this-relate-to-a-person-having-to-approve","How does this relate to a person having to approve?",[174,532,533,534,347],{},"A human gate only counts if you can later show who signed and what they saw. Without a graph, that gate is a popup that forgets. See ",[219,535,537],{"href":536},"what-is-human-in-the-loop-ai","What is human-in-the-loop AI",[333,539,541],{"id":540},"does-the-graph-replace-crm-field-history","Does the graph replace CRM field history?",[174,543,544],{},"No. Field history says the value changed. The graph says which job, which policy version, and which named signer caused the proposal. Keep both. Point; do not duplicate.",[333,546,548],{"id":547},"what-if-the-models-explanation-disagrees-with-the-graph","What if the model’s explanation disagrees with the graph?",[174,550,551],{},"Trust the structure. Fluent “because” text is often written after the fact. The chain of brief, sources, quote, and signature is the operational cause.",[333,553,555],{"id":554},"can-we-store-every-prompt-and-completion","Can we store every prompt and completion?",[174,557,558,559,347],{},"You can. You usually should not. Completeness is reconstructability, not hoarding. Secrets in transcripts become a new breach class. See ",[219,560,561],{"href":275},"What is institutional memory in enterprise AI",[333,563,565],{"id":564},"how-do-permissions-work-on-the-graph","How do permissions work on the graph?",[174,567,568],{},"The same least-privilege instinct as the job. If you could not see the People Ops workstream, you should not query its chain in ordinary language either.",[333,570,572],{"id":571},"is-a-screenshot-of-the-approval-enough","Is a screenshot of the approval enough?",[174,574,575],{},"For a one-off incident, maybe. For a control, no. Screenshots do not join, do not retain by type, and do not survive the laptop.",[333,577,579],{"id":578},"where-does-spend-sit-on-the-chain","Where does spend sit on the chain?",[174,581,582,583,347],{},"Quotes, caps, and stop-on-budget are causal events. “The run did not write because the ceiling was hit” is an answer finance can use. See ",[219,584,346],{"href":345},[333,586,588],{"id":587},"how-is-this-different-from-mlops-experiment-tracking","How is this different from MLOps experiment tracking?",[174,590,591,592,595],{},"MLOps tracks model training and deployment. A lifecycle graph tracks operational work that ",[269,593,594],{},"uses"," models. They stack. They do not substitute.",[192,597,599],{"id":598},"related-reading","Related reading",[174,601,602,603,483,605,607,608,347],{},"If the goal is what the company still knows after people leave, read ",[219,604,561],{"href":275},[219,606,487],{"href":486},". For the science versus operations cut, ",[219,609,327],{"href":326},[192,611,613],{"id":612},"sources","Sources",[197,615,616,622,629],{},[200,617,618],{},[219,619,621],{"href":221,"rel":620},[223],"W3C PROV overview",[200,623,624],{},[219,625,628],{"href":626,"rel":627},"https://www.w3.org/TR/prov-dm/",[223],"W3C PROV data model",[200,630,631],{},[219,632,634],{"href":407,"rel":633},[223],"ICO, AI and data protection",{"title":155,"searchDepth":156,"depth":156,"links":636},[637,638,644,645,659,660],{"id":194,"depth":156,"text":195},{"id":280,"depth":156,"text":281,"children":639},[640,642,643],{"id":335,"depth":641,"text":336},3,{"id":374,"depth":641,"text":375},{"id":413,"depth":641,"text":414},{"id":454,"depth":156,"text":455},{"id":490,"depth":156,"text":491,"children":646},[647,648,649,650,651,652,653,654,655,656,657,658],{"id":494,"depth":641,"text":495},{"id":505,"depth":641,"text":506},{"id":515,"depth":641,"text":516},{"id":522,"depth":641,"text":523},{"id":529,"depth":641,"text":530},{"id":540,"depth":641,"text":541},{"id":547,"depth":641,"text":548},{"id":554,"depth":641,"text":555},{"id":564,"depth":641,"text":565},{"id":571,"depth":641,"text":572},{"id":578,"depth":641,"text":579},{"id":587,"depth":641,"text":588},{"id":598,"depth":156,"text":599},{"id":612,"depth":156,"text":613},"2026-08-17","A lifecycle graph is how a company answers “why did this happen?” after AI is involved — the chain of cause and effect, not a chat log.","/blog/what-is-a-lifecycle-graph",{"title":165,"description":662},"explainer","blog/what-is-a-lifecycle-graph",[665,668,669,670],"lifecycle-graph","institutional-memory","audit","h3JSROHGCfRknyBr-B5YQvfKgkgs4vl4qnJ6vzw_PHE",{"hero":673,"id":675,"title":676,"archived":149,"authors":150,"badge":150,"body":677,"date":150,"department":150,"description":681,"extension":158,"eyebrow":682,"faqHeader":150,"faqs":150,"footerBand":683,"headline":150,"image":150,"industry":150,"jobType":150,"listed":149,"location":150,"navigation":115,"openRoles":150,"pageLayout":150,"path":60,"relatedHeading":689,"seo":690,"series":150,"sitemap":115,"status":150,"stem":691,"subhead":150,"tags":150,"video":150,"whyJoin":150,"workplaceType":150,"__hash__":692},{"filename":674},"u2221455217_Flat_design_of_a_futuristic_minimalist_landscape__5d589295-cdea-4ea9-a262-be766881accf_1.png","content/blog/index.md","Exploring the future of intelligence.",{"type":152,"value":678,"toc":679},[],{"title":155,"searchDepth":156,"depth":156,"links":680},[],"Deep dives into pre-cognitive intelligence, sentient enterprises, and the evolving landscape of AI-driven business transformation.","Latest Research",{"headline":684,"description":685,"primaryLabel":686,"primaryTo":687,"secondaryLabel":688,"secondaryTo":12},"Stay at the frontier.","Subscribe for product updates and new insights.","Subscribe","/newsletter","Explore the platform","More research",{"title":676,"description":681},"blog/index","eK1RCXdDW8nfLSyKRXGB1mJm9FAmhAO6GWXwNKOMNVE",[694,1266],{"id":695,"title":696,"archived":149,"authors":697,"badge":699,"body":700,"date":661,"department":150,"description":1258,"extension":158,"eyebrow":150,"faqHeader":150,"faqs":150,"footerBand":150,"headline":150,"image":150,"industry":150,"jobType":150,"listed":149,"location":150,"navigation":115,"openRoles":150,"pageLayout":150,"path":1259,"relatedHeading":150,"seo":1260,"series":665,"sitemap":115,"status":150,"stem":1261,"subhead":150,"tags":1262,"video":150,"whyJoin":150,"workplaceType":150,"__hash__":1265},"content/blog/what-is-ai-governance.md","What is AI Governance",[698],{"name":168,"to":120},{"label":170},{"type":152,"value":701,"toc":1234},[702,709,712,715,732,735,749,751,844,854,856,865,867,899,902,930,937,939,944,955,960,965,976,978,984,990,996,1004,1010,1018,1037,1039,1042,1049,1063,1073,1075,1079,1087,1091,1094,1098,1105,1109,1112,1116,1119,1123,1130,1134,1137,1141,1144,1148,1151,1155,1158,1162,1167,1171,1174,1176,1187,1189],[174,703,704,705,708],{},"AI governance is the set of rules, ",[178,706,707],{},"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.",[174,710,711],{},"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.",[174,713,714],{},"People use the phrase for three different things, and they get mixed up:",[716,717,718,726,729],"ol",{},[200,719,720,721,347],{},"A public commitment — for example the ",[219,722,725],{"href":723,"rel":724},"https://oecd.ai/en/ai-principles",[223],"OECD AI Principles",[200,727,728],{},"A company committee with a risk register.",[200,730,731],{},"The runtime that actually stops a change to CRM, ERP, or a customer message.",[174,733,734],{},"All three are real. Only the third one would have blocked an unlogged field change that later showed up in a forecast.",[174,736,737,742,743,748],{},[219,738,741],{"href":739,"rel":740},"https://www.gartner.com/en/articles/ai-governance-trism",[223],"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 ",[219,744,747],{"href":745,"rel":746},"https://www.nist.gov/itl/ai-risk-management-framework",[223],"NIST AI Risk Management Framework"," 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.",[192,750,195],{"id":194},[197,752,753,759,774,783,789,795,809,819,831],{},[200,754,755,758],{},[178,756,757],{},"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.",[200,760,761,764,765,768,769,773],{},[178,762,763],{},"Write / write-back."," The AI is allowed to ",[269,766,767],{},"change"," that system, not only draft a suggestion. See ",[219,770,772],{"href":771},"what-is-write-back-governance","What is write-back governance",". At work, a next-step note and an Amount field are not the same risk class.",[200,775,776,779,780,782],{},[178,777,778],{},"Human-in-the-loop."," A person must approve before the job can finish. See ",[219,781,537],{"href":536},". At work, the gate shows the payload in the language of the live system, not a wall of prompt text.",[200,784,785,788],{},[178,786,787],{},"Named signer."," The identity that authorised the change. At work, “someone in the channel clicked yes” is not a signer.",[200,790,791,794],{},[178,792,793],{},"Fail-closed."," Missing approval means nothing happens. Fail-open means the change goes through unless someone happens to stop it.",[200,796,797,800,801,804,805,808],{},[178,798,799],{},"DPIA."," A data-protection impact assessment — thinking through purpose, risk, and personal data ",[269,802,803],{},"before"," you turn a tool loose. ",[219,806,409],{"href":407,"rel":807},[223]," still wants a lawful basis and purpose when the “user” is an AI.",[200,810,811,814,815,347],{},[178,812,813],{},"Shadow AI."," Personal ChatGPT for work because the official path is missing. See ",[219,816,818],{"href":817},"what-is-shadow-ai","What is shadow AI",[200,820,821,824,825,830],{},[178,822,823],{},"Inventory."," A list of where AI actually runs. The ",[219,826,829],{"href":827,"rel":828},"https://www.justice.gov/media/1373026/dl",[223],"US plan described in OMB M-24-10"," puts a named owner and an inventory first, not a PDF.",[200,832,833,836,837,840,841,843],{},[178,834,835],{},"Least privilege."," Only the data and tools required for ",[269,838,839],{},"this"," job. A ",[219,842,487],{"href":486}," is how that instinct becomes a company object.",[174,845,846,847,850,851,853],{},"Model safety is adjacent and different. Safety is about what a model will say in the abstract. Enterprise governance is about what ",[269,848,849],{},"your"," people and tools may do with ",[269,852,849],{}," systems and data. You can have a carefully aligned model and still have ungoverned CRM writes.",[192,855,281],{"id":280},[174,857,858,859,864],{},"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. ",[219,860,863],{"href":861,"rel":862},"https://www.cbc.ca/news/canada/british-columbia/air-canada-chatbot-lawsuit-1.7116416",[223],"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.",[174,866,290],{},[197,868,869,875,881,893],{},[200,870,871,874],{},[178,872,873],{},"Own a number."," Forecasts and close packs inherit whatever changed.",[200,876,877,880],{},[178,878,879],{},"Own a customer relationship."," Model output that becomes a commitment is still the company’s commitment.",[200,882,883,886,887,892],{},[178,884,885],{},"Own risk or legal."," Privacy law does not pause for a chatbot. ",[219,888,891],{"href":889,"rel":890},"https://eur-lex.europa.eu/eli/reg/2016/679/oj",[223],"GDPR"," still applies to purpose, minimisation, and erasure.",[200,894,895,898],{},[178,896,897],{},"Are asked “who is in charge of AI here?”"," An inventory and a named owner beat a principles slide.",[174,900,901],{},"Good governance in practice is four working rules:",[197,903,904,910,916,922],{},[200,905,906,909],{},[178,907,908],{},"People and rights."," Humans and AI tools are both actors. Roles decide what they may start, see, and sign.",[200,911,912,915],{},[178,913,914],{},"Data at question time."," Purpose and minimisation still apply when an AI is the one looking.",[200,917,918,921],{},[178,919,920],{},"Action rights."," Read-only is a control. Unrestricted tools are an incident waiting for a bad prompt.",[200,923,924,927,928,347],{},[178,925,926],{},"Evidence and spend."," Chat scrollback is not a management system. Uncapped spend is a budget failure and often a security failure. See ",[219,929,346],{"href":345},[174,931,932,933,936],{},"Blocking consumer ChatGPT at the office network, while people use personal phones, is not governance. It is a ",[219,934,935],{"href":817},"shadow AI"," problem with extra steps.",[333,938,336],{"id":335},[174,940,941,943],{},[178,942,341],{}," 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.",[174,945,946,948,949,954],{},[178,947,352],{}," 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, ",[219,950,953],{"href":951,"rel":952},"https://eur-lex.europa.eu/eli/reg/2024/1689/oj",[223],"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.",[174,956,957,959],{},[178,958,358],{}," 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.",[174,961,962,964],{},[178,963,364],{}," 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.",[174,966,967,969,970,975],{},[178,968,370],{}," 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 ",[219,971,974],{"href":972,"rel":973},"https://genai.owasp.org/llm-top-10/",[223],"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.",[333,977,414],{"id":413},[174,979,980,983],{},[178,981,982],{},"Governance as a committee."," Useful for risk registers. Useless if the product can still write.",[174,985,986,989],{},[178,987,988],{},"Governance as model safety."," Refusals on public-web questions do not bind Salesforce.",[174,991,992,995],{},[178,993,994],{},"Governance as a secure web gateway."," Necessary for some paste-out paths. Insufficient for writes, approvals, and causal history.",[174,997,998,1001,1002,347],{},[178,999,1000],{},"Governance as blocking."," Blocks without a sanctioned path train people onto phones. See ",[219,1003,818],{"href":817},[174,1005,1006,1009],{},[178,1007,1008],{},"Theatre."," A checkbox, a prompt that says “ask first,” or an admin toggle the model can ignore.",[174,1011,1012,1013,1017],{},"Good looks like: connectors default to read-only; writes are quoted; a named signer cannot be waived by the model; evidence lives on a ",[219,1014,1016],{"href":1015},"what-is-a-lifecycle-graph","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.",[174,1019,1020,1021,1024,1025,1028,1029,1031,1032,1036],{},"Adjacent concepts: ",[219,1022,1023],{"href":771},"write-back governance"," is the write subset. ",[219,1026,1027],{"href":536},"Human-in-the-loop"," is the gate. ",[219,1030,31],{"href":486}," are the isolation unit. An ",[219,1033,1035],{"href":1034},"what-is-an-enterprise-ai-operating-system","enterprise AI operating system"," is the product shape that makes those rules the default path.",[192,1038,455],{"id":454},[174,1040,1041],{},"Nimbus treats governance as how work is released, not as a sidecar policy engine.",[174,1043,1044,1045,1048],{},"Connectors — secure links to live systems — default to ",[178,1046,1047],{},"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.",[174,1050,1051,1052,1054,1055,1057,1058,1062],{},"Scope is the ",[219,1053,487],{"href":486},": one job, with the playbooks, systems, teams, and budget that belong to that job. Evidence is the ",[219,1056,23],{"href":1015},". The ",[219,1059,1061],{"href":1060},"what-is-a-company-wiki-for-ai-agents","company wiki"," is the asserted policy the run must cite. Model routing does not bypass the gate.",[174,1064,1065,1066,1068,1069,347],{},"See ",[219,1067,39],{"href":40},". For scoring vendors: ",[219,1070,1072],{"href":1071},"how-to-evaluate-ai-governance-platforms","How to evaluate AI governance platforms",[192,1074,491],{"id":490},[333,1076,1078],{"id":1077},"is-ai-governance-the-same-as-making-the-model-safe","Is AI governance the same as making the model “safe”?",[174,1080,1081,1082,850,1084,1086],{},"No. Model safety is about what the model will say in the abstract. Enterprise governance is about what ",[269,1083,849],{},[269,1085,849],{}," systems and data.",[333,1088,1090],{"id":1089},"can-we-rely-on-the-secure-web-gateway","Can we rely on the secure web gateway?",[174,1092,1093],{},"Network controls help with paste-out. They do not quote a CRM change, bind an approver, or store a causal history. Use both.",[333,1095,1097],{"id":1096},"must-a-person-always-approve","Must a person always approve?",[174,1099,1100,1101,1104],{},"For many operational writes, yes. For read-only analysis, maybe not. The mistake is calling a system “human-approved” because a human ",[269,1102,1103],{},"could"," look, while changes proceed on model initiative.",[333,1106,1108],{"id":1107},"do-the-oecd-ai-principles-require-a-specific-product","Do the OECD AI Principles require a specific product?",[174,1110,1111],{},"No. They are a public commitment. A product can make evidence cheaper to produce. The commitment does not implement a gate.",[333,1113,1115],{"id":1114},"is-a-dpia-enough-to-go-live","Is a DPIA enough to go live?",[174,1117,1118],{},"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.",[333,1120,1122],{"id":1121},"does-blocking-chatgpt-count-as-governance","Does blocking ChatGPT count as governance?",[174,1124,1125,1126,1129],{},"It is a network control. Without a sanctioned path that can see the right files, people use personal phones. Blocking can tighten ",[269,1127,1128],{},"after"," substitution exists.",[333,1131,1133],{"id":1132},"how-is-this-different-from-it-change-management","How is this different from IT change management?",[174,1135,1136],{},"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.",[333,1138,1140],{"id":1139},"who-should-be-the-named-owner-of-ai","Who should be the named owner of AI?",[174,1142,1143],{},"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.",[333,1145,1147],{"id":1146},"can-we-govern-only-customer-facing-chatbots-and-ignore-internal-copilots","Can we govern only customer-facing chatbots and ignore internal copilots?",[174,1149,1150],{},"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.",[333,1152,1154],{"id":1153},"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?",[174,1156,1157],{},"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.",[333,1159,1161],{"id":1160},"where-does-spend-fit","Where does spend fit?",[174,1163,1164,1165,347],{},"Uncapped inference is a control failure. Quotes, ceilings, and attribution by job are governance of a scarce, abusable resource. See ",[219,1166,346],{"href":345},[333,1168,1170],{"id":1169},"is-an-acceptable-use-policy-still-worth-writing","Is an acceptable-use policy still worth writing?",[174,1172,1173],{},"Yes, as communication. No, as enforcement. Write the PDF. Then put the same rules in the product people actually use.",[192,1175,599],{"id":598},[174,1177,1178,1180,1181,1183,1184,347],{},[219,1179,772],{"href":771},", ",[219,1182,818],{"href":817},", and ",[219,1185,1186],{"href":1034},"What is an enterprise AI operating system",[192,1188,613],{"id":612},[197,1190,1191,1197,1202,1207,1212,1218,1223,1229],{},[200,1192,1193],{},[219,1194,1196],{"href":739,"rel":1195},[223],"Gartner, AI governance and TRiSM",[200,1198,1199],{},[219,1200,725],{"href":723,"rel":1201},[223],[200,1203,1204],{},[219,1205,634],{"href":407,"rel":1206},[223],[200,1208,1209],{},[219,1210,747],{"href":745,"rel":1211},[223],[200,1213,1214],{},[219,1215,1217],{"href":861,"rel":1216},[223],"CBC, Air Canada chatbot lawsuit",[200,1219,1220],{},[219,1221,891],{"href":889,"rel":1222},[223],[200,1224,1225],{},[219,1226,1228],{"href":951,"rel":1227},[223],"EU AI Act (Regulation 2024/1689)",[200,1230,1231],{},[219,1232,974],{"href":972,"rel":1233},[223],{"title":155,"searchDepth":156,"depth":156,"links":1235},[1236,1237,1241,1242,1256,1257],{"id":194,"depth":156,"text":195},{"id":280,"depth":156,"text":281,"children":1238},[1239,1240],{"id":335,"depth":641,"text":336},{"id":413,"depth":641,"text":414},{"id":454,"depth":156,"text":455},{"id":490,"depth":156,"text":491,"children":1243},[1244,1245,1246,1247,1248,1249,1250,1251,1252,1253,1254,1255],{"id":1077,"depth":641,"text":1078},{"id":1089,"depth":641,"text":1090},{"id":1096,"depth":641,"text":1097},{"id":1107,"depth":641,"text":1108},{"id":1114,"depth":641,"text":1115},{"id":1121,"depth":641,"text":1122},{"id":1132,"depth":641,"text":1133},{"id":1139,"depth":641,"text":1140},{"id":1146,"depth":641,"text":1147},{"id":1153,"depth":641,"text":1154},{"id":1160,"depth":641,"text":1161},{"id":1169,"depth":641,"text":1170},{"id":598,"depth":156,"text":599},{"id":612,"depth":156,"text":613},"AI governance is 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.","/blog/what-is-ai-governance",{"title":696,"description":1258},"blog/what-is-ai-governance",[665,1263,1264,670],"governance","compliance","8Gc6YA0kBilCFVsDbCgM_sj0gjR1GSQazeF0iy3Fot8",{"id":1267,"title":1268,"archived":149,"authors":1269,"badge":1271,"body":1272,"date":661,"department":150,"description":1705,"extension":158,"eyebrow":150,"faqHeader":150,"faqs":150,"footerBand":150,"headline":150,"image":150,"industry":150,"jobType":150,"listed":149,"location":150,"navigation":115,"openRoles":150,"pageLayout":150,"path":1706,"relatedHeading":150,"seo":1707,"series":665,"sitemap":115,"status":150,"stem":1708,"subhead":150,"tags":1709,"video":150,"whyJoin":150,"workplaceType":150,"__hash__":1713},"content/blog/what-is-a-company-wiki-for-ai-agents.md","What is a Company Wiki for AI Agents",[1270],{"name":168,"to":120},{"label":170},{"type":152,"value":1273,"toc":1680},[1274,1281,1284,1291,1294,1296,1299,1330,1333,1381,1384,1386,1392,1395,1425,1427,1439,1448,1453,1458,1467,1469,1476,1483,1490,1494,1497,1504,1507,1519,1521,1527,1534,1537,1548,1550,1554,1561,1565,1568,1572,1575,1579,1582,1586,1596,1600,1603,1607,1610,1614,1617,1621,1628,1632,1638,1642,1649,1653,1656,1658,1666,1668],[174,1275,1276,1277,1280],{},"A company wiki for AI agents is the ",[178,1278,1279],{},"official playbook the AI must follow",": owned, versioned, and scoped — not a pile of old Drive files that search might find.",[174,1282,1283],{},"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.",[174,1285,1286,1287,1290],{},"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 ",[269,1288,1289],{},"sounds"," closest to the question. That is not policy. That is folklore with a search box.",[174,1292,1293],{},"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.",[192,1295,195],{"id":194},[174,1297,1298],{},"Keep three kinds of text apart:",[197,1300,1301,1307,1321],{},[200,1302,1303,1306],{},[178,1304,1305],{},"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.",[200,1308,1309,1312,1313,1316,1317,1320],{},[178,1310,1311],{},"Retrieved."," What exists in systems. Possibly stale or contradictory. That is ",[219,1314,1315],{"href":265},"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 ",[269,1318,1319],{},"records",". They are not automatically the rule you want the agent to follow next.",[200,1322,1323,1326,1327,1329],{},[178,1324,1325],{},"Decided."," What we already approved in a run, stored on the ",[219,1328,1016],{"href":1015},". 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.",[174,1331,1332],{},"Other terms you will hear in vendor decks and internal Slack, and how they actually show up:",[197,1334,1335,1341,1352,1358,1364,1370,1376],{},[200,1336,1337,1340],{},[178,1338,1339],{},"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.”",[200,1342,1343,1346,1347,1351],{},[178,1344,1345],{},"Citation."," The answer names the page and version — ",[1348,1349,1350],"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.",[200,1353,1354,1357],{},[178,1355,1356],{},"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.",[200,1359,1360,1363],{},[178,1361,1362],{},"Authority marker."," Labels such as policy, draft, archive, and local exception. Drafts must not load as binding context.",[200,1365,1366,1369],{},[178,1367,1368],{},"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.",[200,1371,1372,1375],{},[178,1373,1374],{},"Review cadence."," A date when the page is re-checked. Silence becomes folklore.",[200,1377,1378,1380],{},[178,1379,254],{}," Which jobs may load this page. People-ops rules are not in the go-to-market context by default.",[174,1382,1383],{},"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.",[192,1385,281],{"id":280},[174,1387,1388,1389,1391],{},"When official policy is unusable, employees ask consumer models to invent policy. See ",[219,1390,818],{"href":817},". The unofficial tool will synthesise a refund rule or a customer commitment from whatever was pasted. The company still owns the result.",[174,1393,1394],{},"It affects you if:",[197,1396,1397,1403,1411,1417],{},[200,1398,1399,1402],{},[178,1400,1401],{},"Numbers in playbooks disagree"," with numbers in CRM, and nobody can say which is official.",[200,1404,1405,1408,1409,347],{},[178,1406,1407],{},"People leave."," Tacit knowledge — the hallway version of the rule — leaves with them. See ",[219,1410,561],{"href":275},[200,1412,1413,1416],{},[178,1414,1415],{},"Legal or finance must cite a version",", not a vibe.",[200,1418,1419,1422,1423,347],{},[178,1420,1421],{},"Agents can propose writes."," A model that can change CRM without a binding playbook is improvising in production. See ",[219,1424,772],{"href":771},[333,1426,336],{"id":335},[174,1428,1429,1431,1432,1434,1435,1438],{},[178,1430,341],{}," 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 ",[269,1433,803],{}," 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 ",[219,1436,1437],{"href":345},"token economics"," inflate without improving the artefact.",[174,1440,1441,1443,1444,1447],{},[178,1442,352],{}," 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 ",[269,1445,1446],{},"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.",[174,1449,1450,1452],{},[178,1451,358],{}," Runbooks, escalation thresholds, and supplier exception rules need to be loadable as the current procedure, not as the 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.",[174,1454,1455,1457],{},[178,1456,364],{}," Discount floors, win/loss 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.",[174,1459,1460,1462,1463,347],{},[178,1461,370],{}," 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 ",[219,1464,1466],{"href":1465},"what-is-ai-governance","What is AI governance",[333,1468,414],{"id":413},[174,1470,1471,1472,1475],{},"The common failure is treating ",[178,1473,1474],{},"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.",[174,1477,1478,1479,1482],{},"The second failure is the ",[178,1480,1481],{},"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.",[174,1484,1485,1486,1489],{},"The third failure is the ",[178,1487,1488],{},"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.",[333,1491,1493],{"id":1492},"what-good-looks-like-versus-what-fails","What good looks like versus what fails",[174,1495,1496],{},"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.",[174,1498,1499,1500,1503],{},"A good wiki is also ",[178,1501,1502],{},"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.",[174,1505,1506],{},"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.”",[174,1508,1509,1510,1512,1513,1515,1516,1518],{},"Adjacent ideas: retrieval without assertion is ",[219,1511,1315],{"href":265},". Decisions without a playbook are a ",[219,1514,1016],{"href":1015}," with nothing to cite. A job that loads the wrong vault is a ",[219,1517,487],{"href":486}," with the wrong attachments.",[192,1520,455],{"id":454},[174,1522,1523,1524,1526],{},"The ",[178,1525,27],{}," 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.",[174,1528,1529,1530,1533],{},"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 ",[269,1531,1532],{},"against",". When a write is proposed, the named signer should see the playbook version the draft claims to follow.",[174,1535,1536],{},"Perception can ask what the current playbook says, and which run last cited it.",[174,1538,1065,1539,1542,1543,1545,1546,347],{},[219,1540,1541],{"href":28},"Wiki",". For the job that loads a subset of pages, see ",[219,1544,31],{"href":32},". For the chain that records which version was used, see ",[219,1547,23],{"href":24},[192,1549,491],{"id":490},[333,1551,1553],{"id":1552},"isnt-this-just-confluence","Isn’t this just Confluence?",[174,1555,1556,1557,1560],{},"Confluence is a human wiki. An agent wiki is a ",[269,1558,1559],{},"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.”",[333,1562,1564],{"id":1563},"cant-we-just-search-drive","Can’t we just search Drive?",[174,1566,1567],{},"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.",[333,1569,1571],{"id":1570},"what-if-the-wiki-is-wrong","What if the wiki is wrong?",[174,1573,1574],{},"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.”",[333,1576,1578],{"id":1577},"how-is-this-different-from-a-custom-gpts-instructions","How is this different from a custom GPT’s instructions?",[174,1580,1581],{},"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.",[333,1583,1585],{"id":1584},"do-we-need-a-wiki-if-we-already-have-rag","Do we need a wiki if we already have RAG?",[174,1587,1588,1589,1592,1593,347],{},"Yes, if agents will act. RAG reduces invention on ",[269,1590,1591],{},"existing"," files. It does not mark which file is in force. Without assertion, retrieval-augmented generation is retrieval-augmented folklore. See ",[219,1594,1595],{"href":265},"What is enterprise RAG",[333,1597,1599],{"id":1598},"who-should-own-wiki-pages","Who should own wiki pages?",[174,1601,1602],{},"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.",[333,1604,1606],{"id":1605},"how-often-should-pages-be-reviewed","How often should pages be reviewed?",[174,1608,1609],{},"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.",[333,1611,1613],{"id":1612},"what-should-we-put-in-tables-versus-prose","What should we put in tables versus prose?",[174,1615,1616],{},"Numbers, thresholds, codes, and “never / always” lists belong in tables. Narrative belongs in prose. Agents quote tables more reliably than they extract a number buried in a paragraph.",[333,1618,1620],{"id":1619},"can-one-wiki-serve-the-whole-company","Can one wiki serve the whole company?",[174,1622,1623,1624,1627],{},"One ",[269,1625,1626],{},"product",", many vaults. A single unscoped corpus recreates the god workspace. Finance close pages and recruiting pages should not share a default context.",[333,1629,1631],{"id":1630},"how-do-exceptions-work-without-rewriting-the-playbook","How do exceptions work without rewriting the playbook?",[174,1633,1634,1635,347],{},"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 ",[219,1636,1637],{"href":1015},"What is a lifecycle graph",[333,1639,1641],{"id":1640},"will-a-better-model-make-the-wiki-unnecessary","Will a better model make the wiki unnecessary?",[174,1643,1644,1645,347],{},"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 ",[219,1646,1648],{"href":1647},"what-is-model-routing","What is model routing",[333,1650,1652],{"id":1651},"how-does-this-relate-to-access-control","How does this relate to access control?",[174,1654,1655],{},"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.",[192,1657,599],{"id":598},[174,1659,1660,1180,1662,1183,1664,347],{},[219,1661,1595],{"href":265},[219,1663,561],{"href":275},[219,1665,1466],{"href":1465},[192,1667,613],{"id":612},[197,1669,1670,1675],{},[200,1671,1672],{},[219,1673,634],{"href":407,"rel":1674},[223],[200,1676,1677],{},[219,1678,891],{"href":889,"rel":1679},[223],{"title":155,"searchDepth":156,"depth":156,"links":1681},[1682,1683,1688,1689,1703,1704],{"id":194,"depth":156,"text":195},{"id":280,"depth":156,"text":281,"children":1684},[1685,1686,1687],{"id":335,"depth":641,"text":336},{"id":413,"depth":641,"text":414},{"id":1492,"depth":641,"text":1493},{"id":454,"depth":156,"text":455},{"id":490,"depth":156,"text":491,"children":1690},[1691,1692,1693,1694,1695,1696,1697,1698,1699,1700,1701,1702],{"id":1552,"depth":641,"text":1553},{"id":1563,"depth":641,"text":1564},{"id":1570,"depth":641,"text":1571},{"id":1577,"depth":641,"text":1578},{"id":1584,"depth":641,"text":1585},{"id":1598,"depth":641,"text":1599},{"id":1605,"depth":641,"text":1606},{"id":1612,"depth":641,"text":1613},{"id":1619,"depth":641,"text":1620},{"id":1630,"depth":641,"text":1631},{"id":1640,"depth":641,"text":1641},{"id":1651,"depth":641,"text":1652},{"id":598,"depth":156,"text":599},{"id":612,"depth":156,"text":613},"A company wiki for AI agents is the official playbook the AI must follow — versioned, owned, and scoped — not a pile of old Drive files the search might find.","/blog/what-is-a-company-wiki-for-ai-agents",{"title":1268,"description":1705},"blog/what-is-a-company-wiki-for-ai-agents",[665,1710,1711,1712],"wiki","agents","knowledge","lsA-Idy_JOpux3bMMMzFoUJ-4x86_CXZqPG7x0pR-eE",{"enabled":149,"message":1715,"linkLabel":78,"linkHref":79,"id":1716,"title":1717,"archived":149,"authors":150,"badge":150,"body":1718,"date":150,"department":150,"description":155,"extension":158,"eyebrow":150,"faqHeader":150,"faqs":150,"footerBand":150,"headline":150,"image":150,"industry":150,"jobType":150,"listed":149,"location":150,"navigation":115,"openRoles":150,"pageLayout":150,"path":1722,"relatedHeading":150,"seo":1723,"series":150,"sitemap":115,"status":150,"stem":1724,"subhead":150,"tags":150,"video":150,"whyJoin":150,"workplaceType":150,"__hash__":1725},"We're hiring! Join the team building the Sentient Enterprise.","content/shared/hiring.md","Hiring banner",{"type":152,"value":1719,"toc":1720},[],{"title":155,"searchDepth":156,"depth":156,"links":1721},[],"/shared/hiring",{"title":1717,"description":155},"shared/hiring","-6bioYD7lKYokGUVU3ff4hHTvB-sDyOMuCptKHnojfk",{"fold":1727,"id":1731,"title":1732,"archived":149,"authors":150,"badge":150,"body":1733,"date":150,"department":150,"description":155,"extension":158,"eyebrow":150,"faqHeader":150,"faqs":150,"footerBand":1737,"headline":150,"image":150,"industry":150,"jobType":150,"listed":149,"location":150,"navigation":115,"openRoles":150,"pageLayout":150,"path":1741,"relatedHeading":150,"seo":1742,"series":150,"sitemap":115,"status":150,"stem":1743,"subhead":150,"tags":150,"video":150,"whyJoin":150,"workplaceType":150,"__hash__":1744},{"headline":1728,"description":1729,"primaryLabel":8,"primaryTo":1730,"secondaryLabel":688,"secondaryTo":12},"Run frontier AI your business actually owns.","Governed agent swarms, 2,000+ integrations, and a knowledge graph that stays inside your walls. Free 7-day trial.","/checkout","content/shared/cta.md","Site CTAs",{"type":152,"value":1734,"toc":1735},[],{"title":155,"searchDepth":156,"depth":156,"links":1736},[],{"headline":1738,"description":1739,"primaryLabel":8,"primaryTo":1730,"secondaryLabel":1740,"secondaryTo":84},"See what governed AI looks like on your stack.","Connect your tools, run a workstream, and keep every decision on your ledger - free for 7 days.","Talk to our team","/shared/cta",{"title":1732,"description":155},"shared/cta","YHK6Fb8AvCPR1zZq7R_xiXUG0hwhP5UxHA8Ix52JQp4",1787194072231]