[{"data":1,"prerenderedAt":1613},["ShallowReactive",2],{"site-nav-content":3,"blog:/blog/what-is-model-context-protocol":163,"blog-index-copy":685,"blog:/blog/what-is-model-context-protocol:surround":706,"hiring-banner-content":1582,"site-cta-content":1594},{"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":675,"department":150,"description":676,"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":677,"relatedHeading":150,"seo":678,"series":679,"sitemap":115,"status":150,"stem":680,"subhead":150,"tags":681,"video":150,"whyJoin":150,"workplaceType":150,"__hash__":684},"content/blog/what-is-model-context-protocol.md","What is Model Context Protocol",[167],{"name":168,"to":120},"Nimbus Research",{"label":170},"Explainer",{"type":152,"value":172,"toc":650},[173,177,204,211,220,223,228,284,296,300,303,310,331,334,348,351,356,367,373,379,385,397,401,407,413,419,429,439,448,451,474,481,485,494,497,505,509,513,516,520,523,527,533,537,540,544,547,551,554,558,561,565,568,572,577,581,586,590,598,602,605,609,616,620],[174,175,176],"p",{},"USB did not create a data-governance programme. A common plug let keyboards, cameras, and drives talk to any computer. It did not decide who may copy the finance drive, or whether a change to the ledger needs a signer.",[174,178,179,183,184,191,192,197,198,203],{},[180,181,182],"strong",{},"Model Context Protocol (MCP)"," is the same kind of open standard for AI. ",[185,186,190],"a",{"href":187,"rel":188},"https://www.anthropic.com/news/model-context-protocol",[189],"nofollow","Anthropic announced it"," as a ",[185,193,196],{"href":194,"rel":195},"https://modelcontextprotocol.io/docs/2026-07-28/getting-started/intro",[189],"common plug"," so AI apps can use the same tools and files, instead of every vendor inventing a one-off connection. The ",[185,199,202],{"href":200,"rel":201},"https://modelcontextprotocol.io/specification/2025-11-25/index",[189],"specification"," standardises how a host calls tools and reads resources.",[174,205,206,207,210],{},"In one sentence: MCP is ",[180,208,209],{},"plumbing, not a company strategy",". It does not decide who may update Salesforce.",[174,212,213,214,219],{},"Developers already know this pattern from the ",[185,215,218],{"href":216,"rel":217},"https://microsoft.github.io/language-server-protocol/",[189],"Language Server Protocol",": one language server, many editors, instead of rewriting autocomplete for every IDE. MCP is that idea for tools an AI can call. LSP made language servers interchangeable. It did not make every language server a safe place for customer lists.",[174,221,222],{},"Before this standard, every AI product invented its own way to “use a tool.” Teams spent months redoing the same wiring. That cost was real. So is the over-read: “we support MCP” is not “we have enterprise governance.”",[224,225,227],"h2",{"id":226},"words-youll-hear","Words you’ll hear",[229,230,231,238,244,250,256,262,273],"ul",{},[232,233,234,237],"li",{},[180,235,236],{},"Protocol / standard."," Agreed wiring so products can interoperate. At work, this is the USB cable, not the access-control list on the share.",[232,239,240,243],{},[180,241,242],{},"MCP server / helper."," A small programme that says “here are the actions I can take, and here are the files I can show you.” At work, a helper that searches Drive as a superuser is still a superuser.",[232,245,246,249],{},[180,247,248],{},"Host / client."," The AI application that calls the helper. At work, several hosts can speak MCP and still have completely different write gates — or none.",[232,251,252,255],{},[180,253,254],{},"Tool call."," The AI asking that helper to search a folder, look up a ticket, post a message, or query a database.",[232,257,258,261],{},[180,259,260],{},"Resource."," A file or record the helper can expose for reading.",[232,263,264,267,268,272],{},[180,265,266],{},"Connector (Nimbus)."," A supported, company-controlled integration to a live system — OAuth, scoped to the job, read-only by default. That is the operator-facing story. MCP may sit at a developer edge. It is not a substitute for connectors plus ",[185,269,271],{"href":270},"what-is-ai-governance","governance",".",[232,274,275,278,279,283],{},[180,276,277],{},"Least privilege."," Which tools this ",[185,280,282],{"href":281},"what-is-an-ai-workstream","workstream"," may call. Importing every available helper is how a demo becomes one actor with every production login.",[174,285,286,287,291,292,272],{},"A tool call can still change production data. The protocol will happily pass that change along. The company still has to decide whether that is allowed. Fail-closed writes, named signers, and quoted payloads live ",[288,289,290],"em",{},"above"," the plug. See ",[185,293,295],{"href":294},"what-is-write-back-governance","What is write-back governance",[224,297,299],{"id":298},"why-you-should-care","Why you should care",[174,301,302],{},"The plug is useful. It is also easy to over-read.",[174,304,305,306,309],{},"MCP does ",[180,307,308],{},"not"," decide:",[229,311,312,315,318,321,324],{},[232,313,314],{},"whose login is used",[232,316,317],{},"whether the AI may only read, or also change a live system",[232,319,320],{},"who must approve a change",[232,322,323],{},"how the company remembers what happened",[232,325,326,327],{},"which model is used for the step — see ",[185,328,330],{"href":329},"what-is-model-routing","What is model routing",[174,332,333],{},"Choosing a model is choosing a brain. This standard is choosing hands. A cheap model with dangerous tools is worse than a strong model with none. Decide them separately.",[174,335,336,337,341,342,347],{},"An ",[185,338,340],{"href":339},"what-is-an-agentic-workflow","agentic workflow"," that imports every available tool is a confused workflow. Plumbing is not a stop condition. ",[185,343,346],{"href":344,"rel":345},"https://www.anthropic.com/engineering/building-effective-agents",[189],"Anthropic’s note on building effective agents"," is about bounding tools and stops, not about collecting helpers.",[174,349,350],{},"It affects you if a vendor says “we support MCP” and you hear “we have enterprise governance.” Those are different sentences.",[352,353,355],"h3",{"id":354},"what-changes-by-role","What changes by role",[174,357,358,361,362,366],{},[180,359,360],{},"Finance."," A helper that can post a journal is a write path, protocol or not. Finance should ask whether the host quotes the payload and requires a named signer, not whether the wiring is MCP. Spend also sits above the plug: tool loops can burn ",[185,363,365],{"href":364},"what-is-ai-token-economics","NTUs"," without a ceiling.",[174,368,369,372],{},[180,370,371],{},"Legal."," Processing agreements, purpose, and customer data in helpers running on laptops. Legal should not treat “open standard” as “safe.” A standard plug does not create a DPIA.",[174,374,375,378],{},[180,376,377],{},"Operations."," Bounded tool belts per job. Ops should refuse workflows that attach every helper “for flexibility,” and should keep human wait and budget as stops regardless of how tools are wired.",[174,380,381,384],{},[180,382,383],{},"Go-to-market."," Faster wiring to CRM and Drive can be good — if the connector is still read-only by default. GTM should not confuse a demo that updates an opportunity via MCP with a governed release.",[174,386,387,390,391,396],{},[180,388,389],{},"Security."," This is the sharp edge. Helpers run with some identity. Superuser search is still superuser search. Prompt injection can trick a model into requesting a tool call; the ",[185,392,395],{"href":393,"rel":394},"https://genai.owasp.org/llm-top-10/",[189],"OWASP Top 10 for LLM applications"," is the relevant list. The protocol will not save you. Least privilege, read-only defaults, and fail-closed writes will.",[352,398,400],{"id":399},"what-people-get-wrong","What people get wrong",[174,402,403,406],{},[180,404,405],{},"MCP as governance."," Wiring is not a named signer.",[174,408,409,412],{},[180,410,411],{},"MCP as the Salesforce strategy."," You still need identity, read versus write, an approver, and a record.",[174,414,415,418],{},[180,416,417],{},"Refusing products that do not speak MCP."," Interoperable tools are a plus. Absence of MCP is not absence of a connector. Presence of MCP is not presence of governance.",[174,420,421,424,425,428],{},[180,422,423],{},"Replacing the integration platform."," MCP standardises how an AI ",[288,426,427],{},"talks"," to a helper. Your identity, iPaaS, and change-control stack still have to exist.",[174,430,431,434,435,272],{},[180,432,433],{},"Assuming retrieval will respect permissions."," Only if the helper is built that way. See ",[185,436,438],{"href":437},"what-is-enterprise-rag","What is enterprise RAG",[174,440,441,444,445,447],{},[180,442,443],{},"Collecting every server."," A large tool belt is a confused ",[185,446,340],{"href":339}," and a larger attack surface.",[174,449,450],{},"Good looks like: MCP where it reduces duplicate wiring; operator-facing connectors that stay scoped, encrypted, and read-only by default; writes only after sign-off; no belief that the spec implemented your control framework. Failure looks like a laptop running a superuser helper pointed at production.",[174,452,453,454,457,458,460,461,464,465,469,470,473],{},"Think of the stack in layers, or you will buy the wrong layer. MCP is how a host talks to a helper. A ",[185,455,456],{"href":51},"connector"," is how operators attach a live system to a ",[185,459,282],{"href":281}," with OAuth and a read-only default. ",[185,462,463],{"href":294},"Write-back governance"," is whether a tool call that mutates production is allowed to execute. The ",[185,466,468],{"href":467},"what-is-a-lifecycle-graph","lifecycle graph"," is whether you can still explain the call next quarter. ",[185,471,472],{"href":329},"Model routing"," is which brain issued the call. None of those jobs is in the spec, and that is fine — specs should stay thin. Trouble starts when a thin spec is sold as the thick programme.",[174,475,336,476,480],{},[185,477,479],{"href":478},"what-is-an-enterprise-ai-operating-system","enterprise AI operating system"," sits above plumbing the way an OS sits above USB: isolation, permissions, I/O policy, and state. USB made accessories interchangeable. It did not decide who may format the finance drive.",[224,482,484],{"id":483},"how-this-shows-up-in-nimbus","How this shows up in Nimbus",[174,486,487,488,491,492,272],{},"Nimbus’s operator-facing integrations are ",[180,489,490],{},"connectors",": scoped per workstream, encrypted per tenant, read-only by default. Action connectors write only after human sign-off. See ",[185,493,50],{"href":51},[174,495,496],{},"MCP can be useful at developer edges. It is not the product’s answer to “who may change CRM.” Governance, wiki, and the Lifecycle Graph still sit above any plug.",[174,498,499,500,502,503,272],{},"See ",[185,501,39],{"href":40}," and ",[185,504,31],{"href":32},[224,506,508],{"id":507},"questions-people-actually-ask","Questions people actually ask",[352,510,512],{"id":511},"is-mcp-how-we-should-connect-salesforce","Is MCP how we should connect Salesforce?",[174,514,515],{},"Not by itself. You still need identity, read vs write rights, an approver, and a record. A standard plug does not provide those.",[352,517,519],{"id":518},"should-we-refuse-products-that-dont-speak-mcp","Should we refuse products that don’t speak MCP?",[174,521,522],{},"No. Interoperable tools are a plus. Absence of MCP is not absence of a connector. Presence of MCP is not presence of governance.",[352,524,526],{"id":525},"does-mcp-replace-our-integration-platform","Does MCP replace our integration platform?",[174,528,529,530,532],{},"No. It standardises how an AI ",[288,531,427],{}," to a helper. Your integration, identity, and change-control stack still has to exist.",[352,534,536],{"id":535},"can-mcp-make-retrieval-respect-permissions","Can MCP make retrieval respect permissions?",[174,538,539],{},"Only if the helper is built that way. A tool that searches Drive as a superuser is still a superuser.",[352,541,543],{"id":542},"is-mcp-the-same-as-a-nimbus-connector","Is MCP the same as a Nimbus connector?",[174,545,546],{},"No. A connector is the operator-facing, company-controlled integration: OAuth, workstream scope, read-only default. MCP is a developer wiring standard that might sit at an edge.",[352,548,550],{"id":549},"does-the-spec-require-fail-closed-writes","Does the spec require fail-closed writes?",[174,552,553],{},"No. The spec does not require a quoted Salesforce payload, a named approver, or a fail-closed write. Those are product and policy choices.",[352,555,557],{"id":556},"how-does-this-relate-to-usb-and-lsp","How does this relate to USB and LSP?",[174,559,560],{},"USB and LSP are the right analogies: interoperability of accessories and language servers. Neither is an access-control programme. Do not buy MCP as if it were.",[352,562,564],{"id":563},"can-we-let-every-agent-team-install-their-own-mcp-servers","Can we let every agent team install their own MCP servers?",[174,566,567],{},"That is how you get overlapping write rights and no inventory. Treat helpers like production integrations: owners, scope, and a default of read-only.",[352,569,571],{"id":570},"does-mcp-choose-the-model","Does MCP choose the model?",[174,573,574,575,272],{},"No. Routing is which brain you pay for. MCP is which hands that brain can use. Decide them separately. See ",[185,576,330],{"href":329},[352,578,580],{"id":579},"is-we-support-mcp-a-good-rfp-answer-for-governance","Is “we support MCP” a good RFP answer for governance?",[174,582,583,584,272],{},"It is a good answer for tool interoperability. For governance, ask about quotes, named signers, workstream scope, and the ",[185,585,468],{"href":467},[352,587,589],{"id":588},"what-is-the-security-failure-mode","What is the security failure mode?",[174,591,592,593,597],{},"A helper with broad credentials, a host with no gate, and a model tricked into calling ",[594,595,596],"code",{},"update_record",". The protocol did its job. Your company did not.",[352,599,601],{"id":600},"should-customer-facing-bots-get-mcp-tools-to-internal-crm","Should customer-facing bots get MCP tools to internal CRM?",[174,603,604],{},"That is how a public conversation inherits production hands. Scope tools as tightly as you would scope a workstream — usually, do not.",[224,606,608],{"id":607},"related-reading","Related reading",[174,610,611,502,614,272],{},[185,612,613],{"href":339},"What is an agentic workflow",[185,615,295],{"href":294},[224,617,619],{"id":618},"sources","Sources",[229,621,622,628,634,639,644],{},[232,623,624],{},[185,625,627],{"href":187,"rel":626},[189],"Anthropic, Introducing the Model Context Protocol",[232,629,630],{},[185,631,633],{"href":200,"rel":632},[189],"Model Context Protocol specification",[232,635,636],{},[185,637,218],{"href":216,"rel":638},[189],[232,640,641],{},[185,642,395],{"href":393,"rel":643},[189],[232,645,646],{},[185,647,649],{"href":344,"rel":648},[189],"Anthropic, Building effective agents",{"title":155,"searchDepth":156,"depth":156,"links":651},[652,653,658,659,673,674],{"id":226,"depth":156,"text":227},{"id":298,"depth":156,"text":299,"children":654},[655,657],{"id":354,"depth":656,"text":355},3,{"id":399,"depth":656,"text":400},{"id":483,"depth":156,"text":484},{"id":507,"depth":156,"text":508,"children":660},[661,662,663,664,665,666,667,668,669,670,671,672],{"id":511,"depth":656,"text":512},{"id":518,"depth":656,"text":519},{"id":525,"depth":656,"text":526},{"id":535,"depth":656,"text":536},{"id":542,"depth":656,"text":543},{"id":549,"depth":656,"text":550},{"id":556,"depth":656,"text":557},{"id":563,"depth":656,"text":564},{"id":570,"depth":656,"text":571},{"id":579,"depth":656,"text":580},{"id":588,"depth":656,"text":589},{"id":600,"depth":656,"text":601},{"id":607,"depth":156,"text":608},{"id":618,"depth":156,"text":619},"2026-08-17","Model Context Protocol is a common plug so AI apps can use the same tools, like USB for accessories. It is plumbing, not a company strategy, and it does not decide who may update Salesforce.","/blog/what-is-model-context-protocol",{"title":165,"description":676},"explainer","blog/what-is-model-context-protocol",[679,682,490,683],"mcp","tools","pyCsyLW6Bee4HSZZGalnpk2Ef8Ouo6MpiGQykRPKqaI",{"hero":686,"id":688,"title":689,"archived":149,"authors":150,"badge":150,"body":690,"date":150,"department":150,"description":694,"extension":158,"eyebrow":695,"faqHeader":150,"faqs":150,"footerBand":696,"headline":150,"image":150,"industry":150,"jobType":150,"listed":149,"location":150,"navigation":115,"openRoles":150,"pageLayout":150,"path":60,"relatedHeading":702,"seo":703,"series":150,"sitemap":115,"status":150,"stem":704,"subhead":150,"tags":150,"video":150,"whyJoin":150,"workplaceType":150,"__hash__":705},{"filename":687},"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":691,"toc":692},[],{"title":155,"searchDepth":156,"depth":156,"links":693},[],"Deep dives into pre-cognitive intelligence, sentient enterprises, and the evolving landscape of AI-driven business transformation.","Latest Research",{"headline":697,"description":698,"primaryLabel":699,"primaryTo":700,"secondaryLabel":701,"secondaryTo":12},"Stay at the frontier.","Subscribe for product updates and new insights.","Subscribe","/newsletter","Explore the platform","More research",{"title":689,"description":694},"blog/index","eK1RCXdDW8nfLSyKRXGB1mJm9FAmhAO6GWXwNKOMNVE",[707,1127],{"id":708,"title":709,"archived":149,"authors":710,"badge":712,"body":713,"date":675,"department":150,"description":1118,"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":1119,"relatedHeading":150,"seo":1120,"series":679,"sitemap":115,"status":150,"stem":1121,"subhead":150,"tags":1122,"video":150,"whyJoin":150,"workplaceType":150,"__hash__":1126},"content/blog/what-is-model-routing.md","What is Model Routing",[711],{"name":168,"to":120},{"label":170},{"type":152,"value":714,"toc":1094},[715,721,724,727,740,742,803,810,816,818,821,828,850,860,862,867,875,880,885,890,892,898,908,914,920,926,932,940,955,961,963,966,969,979,981,985,988,992,995,999,1002,1006,1009,1013,1016,1020,1026,1030,1033,1037,1040,1044,1047,1051,1056,1060,1063,1067,1070,1072,1078,1080],[174,716,717,718,720],{},"Labs ship a ladder of models: small and cheap, large and expensive. ",[180,719,472],{}," is the policy above that ladder: use a cheaper, faster model for simple steps, and a stronger model only when the task needs it.",[174,722,723],{},"It is not a dropdown labelled “best.” Someone typing “use the best model” for a classify-this-ticket step is how a flagship invoice gets burned on work a compact model could have finished in a second.",[174,725,726],{},"Done well, extract runs on compact models and hard reasoning runs on frontier models. Done poorly, every step hits the most expensive model, spend becomes a surprise, and “we use the best model” becomes an unexamined religion.",[174,728,729,502,734,739],{},[185,730,733],{"href":731,"rel":732},"https://openai.com/api/pricing/",[189],"OpenAI",[185,735,738],{"href":736,"rel":737},"https://www.anthropic.com/pricing",[189],"Anthropic"," publish those ladders in public. The prices change. The shape does not: input, output, and sometimes tools meter differently, and the top rung is many times the compact rung. Finance cannot treat “always flagship” as a quality culture. It is an unbudgeted preference.",[224,741,227],{"id":226},[229,743,744,750,756,762,768,778,788,794],{},[232,745,746,749],{},[180,747,748],{},"Frontier / flagship model."," The strongest (and usually most expensive) model a lab currently sells. At work, this is for judgment: does this clause violate the playbook?",[232,751,752,755],{},[180,753,754],{},"Compact / small model."," Faster and cheaper. Often enough for extract, classify, and summarise. At work, this is “pull the fields from the export.”",[232,757,758,761],{},[180,759,760],{},"Cascade."," Try cheap first; spend the expensive call only when the cheap one is not enough.",[232,763,764,767],{},[180,765,766],{},"Fallback."," If a provider is down or over budget, send the step somewhere else.",[232,769,770,773,774,777],{},[180,771,772],{},"Orchestration."," What steps exist. Different from routing, which is ",[288,775,776],{},"which brain"," each step uses. You can orchestrate a brilliant multi-agent graph and still send every node to the flagship.",[232,779,780,783,784,787],{},[180,781,782],{},"Quality bar."," The reject-rate or rework threshold that decides whether a compact model is good enough on ",[288,785,786],{},"this"," job.",[232,789,790,793],{},[180,791,792],{},"Data residency / data class."," A cheap endpoint may be forbidden for a class of records. Routing is then a compliance table, not only a cost table.",[232,795,796,799,800,272],{},[180,797,798],{},"NTU."," The normalised unit routing is trying to protect. See ",[185,801,802],{"href":364},"What is AI token economics",[174,804,805,806,809],{},"Routing is also not ",[180,807,808],{},"fine-tuning"," (changing a model’s weights). Fine-tuning is a research and ops programme. Routing is an operating policy over models you already buy.",[174,811,812,813,272],{},"Constraints that belong in the route table: data residency, evaluation (you cannot route on vibes), and security (a model with web tools is a different actor than a model with none). Choosing a model is choosing a brain. Choosing tools is choosing hands. Decide them separately. See ",[185,814,165],{"href":815},"what-is-model-context-protocol",[224,817,299],{"id":298},[174,819,820],{},"It affects you if you pay the bill, or if quality on a step is load-bearing.",[174,822,823,824,827],{},"Talk about it as a ",[180,825,826],{},"budget and quality conversation",", not as an ML research project:",[229,829,830,836,842],{},[232,831,832,835],{},[180,833,834],{},"Tag the steps."," Extracting fields from an export is not the same as arguing whether a clause violates policy. If your platform cannot name steps, it cannot route them.",[232,837,838,841],{},[180,839,840],{},"Set a quality bar per step."," “Compact model until human reject rate exceeds X on this job.” Without a bar, routing becomes “always escalate because someone was once unhappy.”",[232,843,844,847,848,272],{},[180,845,846],{},"Keep the gate regardless of model."," A cheap model with a write tool is still a write tool. See ",[185,849,295],{"href":294},[174,851,852,853,856,857,859],{},"A spend ceiling without routing still lets every step hit the flagship until the ceiling kills the run. Routing is how you stay under the ceiling ",[288,854,855],{},"and"," finish the job. See ",[185,858,613],{"href":339}," for why loops without stops dominate the bill.",[352,861,355],{"id":354},[174,863,864,866],{},[180,865,360],{}," Routing is the practical lever on unit cost. Quotes should assume the policy, not the flagship. Finance should ask for approved-updates per NTU, and for evidence that extract steps are not on the top rung. Locking one vendor forever is a pricing and outage choice; routing across providers is a second tape measure.",[174,868,869,871,872,874],{},[180,870,371],{}," Data class and residency can forbid the cheap endpoint. Legal should sit on the route table for those classes, not discover them on an invoice. Customer-facing language may need a stronger model ",[288,873,855],{}," a named signer; routing does not replace the gate.",[174,876,877,879],{},[180,878,377],{}," Steps must be named or you cannot route them. Ops should own fallbacks when a provider is down, and should refuse a single “best” toggle that bypasses the table.",[174,881,882,884],{},[180,883,383],{}," Quality anxiety is strongest here. Measure reject rates on the job. A compact model that extracts next steps may be fine; a compact model that invents a concession is not. Routing on one unhappy anecdote will pin every step to flagship.",[174,886,887,889],{},[180,888,389],{}," A model with browsing or unconstrained tools is a different actor. Routing should not silently add hands. Prompt injection plus a flagship model plus write tools is a worse combination than a compact extract-only step behind a fail-closed gate.",[352,891,400],{"id":399},[174,893,894,897],{},[180,895,896],{},"Always the smartest model."," Use the weakest model that meets the quality bar for that step. Flagship is for judgment, not for labelling.",[174,899,900,903,904,272],{},[180,901,902],{},"Routing as multi-agent."," Several agents is a cast. Routing is which brain each step pays for. See ",[185,905,907],{"href":906},"what-is-multi-agent-ai","What is multi-agent AI",[174,909,910,913],{},[180,911,912],{},"Routing as fine-tuning."," Different programme.",[174,915,916,919],{},[180,917,918],{},"Dropdown labelled “best.”"," That is not a policy. It is a preference that cannot be audited.",[174,921,922,925],{},[180,923,924],{},"Dropping the write gate for a “trusted” model."," Trust the gate. Models change weekly.",[174,927,928,931],{},[180,929,930],{},"Routing on vibes."," One anecdote becomes a permanent escalate. Measure rework.",[174,933,934,935,939],{},"Good looks like: named steps, a route table with cost, quality bar, and data class, cascade where it helps, fallback across providers, gates independent of model, falling unit cost as the ",[185,936,938],{"href":937},"what-is-a-company-wiki-for-ai-agents","wiki"," reduces re-derivation. Failure looks like flagship-everywhere and a board slide about the bill.",[174,941,942,943,946,947,950,951,954],{},"Adjacent ideas worth keeping separate: ",[185,944,945],{"href":364},"token economics"," is quote, cap, and attribute. Routing is which rung of the ladder a named step is allowed to use. ",[185,948,949],{"href":906},"Multi-agent AI"," is how many specialist roles run. You can route a single agent, and you can send a whole agent team to the flagship by mistake. ",[185,952,953],{"href":815},"MCP"," is hands, not brains: do not let a compact extract step inherit a write tool because “the helper was available.”",[174,956,957,958,960],{},"Evaluation has to live on the job, not in a model-arena screenshot. A compact model that extracts fields with a low reject rate is a success even if it would lose a public chatbot bake-off. A flagship model that drafts a concession the wiki forbids is a failure even if it is eloquent. Tie routing reviews to ",[185,959,282],{"href":281}," outcomes — approved writes, rejects, rework — the same way you would review any other operating policy.",[224,962,484],{"id":483},[174,964,965],{},"Nimbus treats routing as an operating decision tied to workstream steps: task type, sensitivity, and cost — not “best everywhere.” Release gates apply regardless of which model drafted the payload.",[174,967,968],{},"NTU quotes and ceilings sit around that policy so operators see a number before they commit. Everyday extract should not consume flagship credits.",[174,970,499,971,973,974,976,977,272],{},[185,972,44],{"href":45},". For the unit of account routing sits inside, ",[185,975,802],{"href":364},". Product: ",[185,978,31],{"href":32},[224,980,508],{"id":507},[352,982,984],{"id":983},"should-we-always-use-the-smartest-model","Should we always use the smartest model?",[174,986,987],{},"No. Use the weakest model that meets the quality bar for that step. Flagship is for judgment, not for labelling.",[352,989,991],{"id":990},"will-routing-make-answers-worse","Will routing make answers worse?",[174,993,994],{},"It can, if you under-route hard steps. Measure rejects and rework on the job. Do not route on a single anecdote.",[352,996,998],{"id":997},"is-this-the-same-as-having-several-agents","Is this the same as having several agents?",[174,1000,1001],{},"No. Several agents is a cast. Routing is which brain each step pays for.",[352,1003,1005],{"id":1004},"can-we-lock-one-vendor-forever","Can we lock one vendor forever?",[174,1007,1008],{},"You can. You will pay for it in price, outages, and lock-in. Routing across providers is how finance keeps a second tape measure.",[352,1010,1012],{"id":1011},"what-is-a-cascade","What is a cascade?",[174,1014,1015],{},"Try the cheap model first. Escalate only when a confidence or quality check says the cheap pass is not enough. It is a tactic inside a policy, not a substitute for naming steps.",[352,1017,1019],{"id":1018},"does-a-better-model-remove-the-need-for-a-wiki","Does a better model remove the need for a wiki?",[174,1021,1022,1023,272],{},"No. Stronger models are better at sounding like policy. Asserted playbooks still win over Drive folklore. See ",[185,1024,1025],{"href":937},"What is a company wiki for AI agents",[352,1027,1029],{"id":1028},"how-do-we-set-a-quality-bar","How do we set a quality bar?",[174,1031,1032],{},"Start with human reject rate and rework on that step. “Compact until rejects exceed X on this workstream” is a bar. “People like the flagship” is not.",[352,1034,1036],{"id":1035},"should-customer-facing-copy-always-use-the-flagship","Should customer-facing copy always use the flagship?",[174,1038,1039],{},"Not always. It should always use a human gate if it asserts a term or a price. Model size does not absorb Air Canada-style risk.",[352,1041,1043],{"id":1042},"what-if-the-cheap-endpoint-is-in-the-wrong-region","What if the cheap endpoint is in the wrong region?",[174,1045,1046],{},"Then it is not cheap; it is forbidden. Put residency in the route table beside price.",[352,1048,1050],{"id":1049},"does-routing-replace-spend-caps","Does routing replace spend caps?",[174,1052,1053,1054,272],{},"No. Caps stop unbounded loops. Routing makes legitimate work affordable under the cap. You want both. See ",[185,1055,802],{"href":364},[352,1057,1059],{"id":1058},"can-the-model-choose-its-own-successor","Can the model choose its own successor?",[174,1061,1062],{},"Letting the model always escalate is how every step becomes flagship. Escalation should be a policy check, not a preference the model expresses.",[352,1064,1066],{"id":1065},"how-does-this-show-up-in-an-rfp","How does this show up in an RFP?",[174,1068,1069],{},"Ask whether steps are named, whether gates apply regardless of model, and whether finance sees a normalised unit. “We use the best models” is not an answer.",[224,1071,608],{"id":607},[174,1073,1074,502,1076,272],{},[185,1075,802],{"href":364},[185,1077,907],{"href":906},[224,1079,619],{"id":618},[229,1081,1082,1088],{},[232,1083,1084],{},[185,1085,1087],{"href":731,"rel":1086},[189],"OpenAI API pricing",[232,1089,1090],{},[185,1091,1093],{"href":736,"rel":1092},[189],"Anthropic pricing",{"title":155,"searchDepth":156,"depth":156,"links":1095},[1096,1097,1101,1102,1116,1117],{"id":226,"depth":156,"text":227},{"id":298,"depth":156,"text":299,"children":1098},[1099,1100],{"id":354,"depth":656,"text":355},{"id":399,"depth":656,"text":400},{"id":483,"depth":156,"text":484},{"id":507,"depth":156,"text":508,"children":1103},[1104,1105,1106,1107,1108,1109,1110,1111,1112,1113,1114,1115],{"id":983,"depth":656,"text":984},{"id":990,"depth":656,"text":991},{"id":997,"depth":656,"text":998},{"id":1004,"depth":656,"text":1005},{"id":1011,"depth":656,"text":1012},{"id":1018,"depth":656,"text":1019},{"id":1028,"depth":656,"text":1029},{"id":1035,"depth":656,"text":1036},{"id":1042,"depth":656,"text":1043},{"id":1049,"depth":656,"text":1050},{"id":1058,"depth":656,"text":1059},{"id":1065,"depth":656,"text":1066},{"id":607,"depth":156,"text":608},{"id":618,"depth":156,"text":619},"Model routing is using a cheaper, faster model for simple steps and a stronger model only when the task needs it — a policy, not a dropdown labelled “best.”","/blog/what-is-model-routing",{"title":709,"description":1118},"blog/what-is-model-routing",[679,1123,1124,1125],"model-routing","token-economics","cost","cT8uO7Fl6bffuYVI7nA4O6i99dYBnBc0RAb25dGjpU0",{"id":1128,"title":1129,"archived":149,"authors":1130,"badge":1132,"body":1133,"date":675,"department":150,"description":1574,"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":1575,"relatedHeading":150,"seo":1576,"series":679,"sitemap":115,"status":150,"stem":1577,"subhead":150,"tags":1578,"video":150,"whyJoin":150,"workplaceType":150,"__hash__":1581},"content/blog/what-is-institutional-memory-in-enterprise-ai.md","What is Institutional Memory in Enterprise AI",[1131],{"name":168,"to":120},{"label":170},{"type":152,"value":1134,"toc":1550},[1135,1138,1141,1144,1153,1155,1158,1210,1213,1216,1242,1249,1251,1254,1257,1277,1280,1282,1289,1300,1305,1310,1319,1321,1327,1333,1339,1349,1355,1361,1364,1385,1398,1401,1403,1406,1412,1428,1430,1434,1437,1441,1444,1448,1453,1457,1460,1464,1469,1473,1476,1480,1483,1487,1490,1494,1497,1501,1504,1508,1511,1515,1518,1520,1529,1531],[174,1136,1137],{},"Institutional memory is what the company still knows after the person who did the work leaves — after the chat vendor changes, after the model version rolls.",[174,1139,1140],{},"Individual memory is a hallway conversation and a personal ChatGPT thread. Company memory is playbooks, signed decisions, and live systems, with access control.",[174,1142,1143],{},"Organisations have always had memory: filing cabinets, shared drives, ERP history, “ask the person who was here last year.” Generative AI created a new amnesia: high-value reasoning happens in disposable threads, on personal accounts, in tools with the wrong retention, or in a vendor’s silo the company cannot query.",[174,1145,1146,1147,1152],{},"This is an evidence topic, not a nostalgia topic. Financial reporting changes have needed reconstructable authorisation for decades. Records-management programmes ask for metadata and assigned responsibility. None of those regimes is satisfied by a personal chat thread the predecessor took with them. ",[185,1148,1151],{"href":1149,"rel":1150},"https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/artificial-intelligence/",[189],"UK ICO guidance on AI and data protection"," still wants purpose and retention thinking when the “user” is a model.",[224,1154,227],{"id":226},[174,1156,1157],{},"Keep four kinds of memory separate on purpose:",[229,1159,1160,1173,1183,1196],{},[232,1161,1162,1165,1166,1169,1170,1172],{},[180,1163,1164],{},"Asserted policy."," What we ",[288,1167,1168],{},"want"," to be true: playbooks, guardrails, approved language. See ",[185,1171,1025],{"href":937},". At work, this is the current discount floor, not last year’s slide.",[232,1174,1175,1178,1179,1182],{},[180,1176,1177],{},"Systems of record."," What ",[288,1180,1181],{},"is"," true in operations: CRM, ERP, HR, the warehouse. These are memory of the business, not of AI work. At work, the opportunity Amount is here. The reason it changed may not be.",[232,1184,1185,1188,1189,1192,1193,1195],{},[180,1186,1187],{},"Decision memory."," Why we ",[288,1190,1191],{},"changed"," something with AI in the loop: briefs, approvals, rejected options, source versions. A ",[185,1194,468],{"href":467},". At work, this is “who signed this exception, against which playbook version.”",[232,1197,1198,1201,1202,1205,1206,1209],{},[180,1199,1200],{},"Retrieved knowledge."," Documents we ",[288,1203,1204],{},"might"," use. That is ",[185,1207,1208],{"href":437},"enterprise RAG",". Lookup without policy and decisions is a search engine, not memory.",[174,1211,1212],{},"If you collapse all four into “one vector store” — a database of text fingerprints used to find similar documents — you get sludge that cannot tell policy from a brainstorm. You also get a new store of sensitive data.",[174,1214,1215],{},"Other terms:",[229,1217,1218,1224,1230,1236],{},[232,1219,1220,1223],{},[180,1221,1222],{},"Hallway knowledge."," The unofficial version of the rule. It leaves with people. Agents will invent a cousin if it is not asserted.",[232,1225,1226,1229],{},[180,1227,1228],{},"Provider logs."," The vendor’s artefact. Not scoped to your jobs, not your access-controlled ledger.",[232,1231,1232,1235],{},[180,1233,1234],{},"Retention."," How long a class of record is kept. Completeness is reconstructability, not hoarding.",[232,1237,1238,1241],{},[180,1239,1240],{},"Perception."," Asking that memory in ordinary language, with permissions still applied.",[174,1243,1244,1248],{},[185,1245,1247],{"href":1246},"what-is-causal-ai-for-operations","Causal operations"," is the “why did this change?” slice of decision memory. It is not a claim about market lift.",[224,1250,299],{"id":298},[174,1252,1253],{},"It affects you the first Monday after someone leaves, and the first time an auditor asks “why is this exception in the CRM when the playbook still says otherwise?”",[174,1255,1256],{},"Three verbs:",[229,1258,1259,1265,1271],{},[232,1260,1261,1264],{},[180,1262,1263],{},"Assert."," Put the rule into a controlled surface. If it only lives in a slide, agents will invent a cousin.",[232,1266,1267,1270],{},[180,1268,1269],{},"Record."," Store the decision chain when AI is in the loop — not every token, the links that let you reconstruct a change.",[232,1272,1273,1276],{},[180,1274,1275],{},"Ask."," Let the next operator query that memory in ordinary language, with permissions still applied.",[174,1278,1279],{},"Causal operations questions (“why did this change?”) need decision memory. Remember outcomes, quotes, approvals, and citations — not every failed token. Wiki needs owners; memory without freshness is last year’s discount floor.",[352,1281,355],{"id":354},[174,1283,1284,1286,1287,272],{},[180,1285,360],{}," Close packs inherit exceptions. Finance needs the playbook version and the signer, not a rumour that “we always accrue this way.” Provider ChatGPT exports are not a SOX-style trail. Spend history also belongs in memory: which job consumed the units, which run stopped on a cap. See ",[185,1288,802],{"href":364},[174,1290,1291,1293,1294,1299],{},[180,1292,371],{}," Discovery, customer commitments, and erasure. Legal should insist that decision memory points at systems of record rather than duplicating them, and that retention is typed. Infinite chat fails a privacy review. ",[185,1295,1298],{"href":1296,"rel":1297},"https://eur-lex.europa.eu/eli/reg/2016/679/oj",[189],"GDPR"," erasure is harder if you indexed everything into sludge.",[174,1301,1302,1304],{},[180,1303,377],{}," Handoffs. The next shift should query “why did this pause?” without reconstructing Slack. Ops should refuse a design that stores every token “because AI” and then cannot delete it.",[174,1306,1307,1309],{},[180,1308,383],{}," Win/loss reasons and discount exceptions walk out the door with account owners. GTM should put asserted playbooks in the wiki and signed exceptions on the graph — not in a personal Claude project.",[174,1311,1312,1314,1315,272],{},[180,1313,389],{}," Memory is a sensitive store. Access control on the graph and wiki is as important as on the CRM. Shadow AI is amnesia by design: the work happened on an account the company cannot query. See ",[185,1316,1318],{"href":1317},"what-is-shadow-ai","What is shadow AI",[352,1320,400],{"id":399},[174,1322,1323,1326],{},[180,1324,1325],{},"CRM as sufficient memory."," CRM remembers the current field. It does not remember which playbook version, which AI run, or which person signed the exception.",[174,1328,1329,1332],{},[180,1330,1331],{},"Exporting ChatGPT threads."," Vendor artefact. Wrong scope. Wrong access control.",[174,1334,1335,1338],{},[180,1336,1337],{},"One vector store for everything."," Policy, brainstorms, tickets, and decisions become an undifferentiated similarity soup.",[174,1340,1341,1344,1345,1348],{},[180,1342,1343],{},"A business knowledge graph as a substitute."," That graph models customers and products. Institutional memory for AI work models ",[180,1346,1347],{},"what we did with models"," — and why.",[174,1350,1351,1354],{},[180,1352,1353],{},"Keeping everything forever."," Hoarding is not completeness. It is a privacy and cost failure.",[174,1356,1357,1360],{},[180,1358,1359],{},"Remembering every token."," Reconstruct the change. Do not archive the model’s scratch reasoning by default.",[174,1362,1363],{},"Good looks like four layers kept apart, owners on wiki pages, a lifecycle graph of decisions, permissions on ask, typed retention, and pointers to live systems. Failure looks like a personal thread, a vendor log, and a vector lake.",[174,1365,1366,1367,1369,1370,1373,1374,1376,1377,1380,1381,1384],{},"Adjacent concepts: ",[185,1368,1208],{"href":437}," is lookup, not memory of what we decided. A ",[185,1371,1372],{"href":937},"company wiki"," is asserted policy, which goes stale without owners. A ",[185,1375,468],{"href":467}," is decision memory of AI-mediated work. ",[185,1378,1379],{"href":1246},"Causal AI for operations"," is the “why did this change?” question that memory should be able to answer. ",[185,1382,1383],{"href":1317},"Shadow AI"," is how memory never starts.",[174,1386,1387,1388,1391,1392,1397],{},"Do not confuse this with a second CRM. Point at the opportunity; do not copy the pipeline. Copies become conflicting official numbers and an erasure problem under ",[185,1389,1298],{"href":1296,"rel":1390},[189],". The ",[185,1393,1396],{"href":1394,"rel":1395},"https://www.w3.org/TR/prov-overview/",[189],"W3C PROV"," idea — entities, activities, agents — is the right instinct for the decision layer: enough structure to reconstruct, not a lake of tokens.",[174,1399,1400],{},"A Monday-morning test is enough. Can the next operator, with the right permissions, find the playbook version, the signed exception, and the live field — without the predecessor’s laptop? If the answer depends on a personal chat vendor, you do not have institutional memory. You have a coincidence that the person has not left yet.",[224,1402,484],{"id":483},[174,1404,1405],{},"Nimbus combines wiki (asserted policy), connectors (systems of record), the Lifecycle Graph (decision memory), Perception (ask), and workstream scoping (who may see what).",[174,1407,1408,1409,1411],{},"Connectors default to read-only, so analysis can be remembered as ",[288,1410,308],{}," having written. Named signers and fail-closed writes make refusals part of memory, not missing events.",[174,1413,1414,1415,1417,1418,1421,1422,1425,1426,272],{},"Product: ",[185,1416,23],{"href":24},", ",[185,1419,1420],{"href":28},"Wiki",", and ",[185,1423,1424],{"href":36},"Perception",". The job boundary is a ",[185,1427,282],{"href":281},[224,1429,508],{"id":507},[352,1431,1433],{"id":1432},"isnt-crm-already-our-memory","Isn’t CRM already our memory?",[174,1435,1436],{},"CRM remembers the current field. It does not remember which playbook version, which AI run, or which person signed the exception.",[352,1438,1440],{"id":1439},"can-we-just-export-chatgpt-threads","Can we just export ChatGPT threads?",[174,1442,1443],{},"Provider logs are the vendor’s artefact. They are not scoped to your jobs, and they are not your access-controlled ledger.",[352,1445,1447],{"id":1446},"how-is-this-different-from-a-knowledge-graph-of-customers-and-products","How is this different from a knowledge graph of customers and products?",[174,1449,1450,1451,1348],{},"That graph models the business domain. Institutional memory for AI work models ",[180,1452,1347],{},[352,1454,1456],{"id":1455},"does-this-mean-storing-everything-forever","Does this mean storing everything forever?",[174,1458,1459],{},"No. Retention follows the type of record. Completeness is reconstructability, not hoarding.",[352,1461,1463],{"id":1462},"how-is-this-different-from-enterprise-rag","How is this different from enterprise RAG?",[174,1465,1466,1467,272],{},"RAG retrieves what exists. Memory of work is what we asserted, what we decided, and what the live system holds. Retrieval without those layers is search. See ",[185,1468,438],{"href":437},[352,1470,1472],{"id":1471},"what-should-we-remember-from-a-run","What should we remember from a run?",[174,1474,1475],{},"The brief, sources (including wiki version), quoted payload, named signer, live-system result, and spend stop if any. Not every failed token, and not secrets in transcripts by default.",[352,1477,1479],{"id":1478},"how-do-we-stop-last-years-policy-living-forever","How do we stop last year’s policy living forever?",[174,1481,1482],{},"Owners and review cadence on the wiki. Archives must not win retrieval against current policy. Freshness is part of memory, not a nice-to-have.",[352,1484,1486],{"id":1485},"can-perception-see-other-departments-decisions","Can Perception see other departments’ decisions?",[174,1488,1489],{},"Only with the same least privilege as the workstream. A go-to-market question should not surface People Ops briefs.",[352,1491,1493],{"id":1492},"is-hallway-knowledge-always-bad","Is hallway knowledge always bad?",[174,1495,1496],{},"It is how work actually happens until you assert it. The failure is leaving it only in hallways once agents are in the loop.",[352,1498,1500],{"id":1499},"how-does-switching-model-vendors-affect-memory","How does switching model vendors affect memory?",[174,1502,1503],{},"If memory lived in the vendor’s chat product, you lost it. If it lived in your wiki, graph, and systems of record, you kept it. That is a buying criterion.",[352,1505,1507],{"id":1506},"where-does-shadow-ai-fit","Where does shadow AI fit?",[174,1509,1510],{},"Personal accounts are institutional amnesia: the company cannot assert, record, or ask. Substitution onto a governed path is how memory starts.",[352,1512,1514],{"id":1513},"do-we-need-a-data-team-to-ask-the-memory","Do we need a data team to ask the memory?",[174,1516,1517],{},"Not if the product has an ordinary-language query surface over the graph and wiki, with permissions. That is Perception in Nimbus. A data team is still right for warehouse metrics.",[224,1519,608],{"id":607},[174,1521,1522,1417,1525,1421,1527,272],{},[185,1523,1524],{"href":467},"What is a lifecycle graph",[185,1526,438],{"href":437},[185,1528,1025],{"href":937},[224,1530,619],{"id":618},[229,1532,1533,1539,1544],{},[232,1534,1535],{},[185,1536,1538],{"href":1149,"rel":1537},[189],"ICO, AI and data protection",[232,1540,1541],{},[185,1542,1298],{"href":1296,"rel":1543},[189],[232,1545,1546],{},[185,1547,1549],{"href":1394,"rel":1548},[189],"W3C PROV overview",{"title":155,"searchDepth":156,"depth":156,"links":1551},[1552,1553,1557,1558,1572,1573],{"id":226,"depth":156,"text":227},{"id":298,"depth":156,"text":299,"children":1554},[1555,1556],{"id":354,"depth":656,"text":355},{"id":399,"depth":656,"text":400},{"id":483,"depth":156,"text":484},{"id":507,"depth":156,"text":508,"children":1559},[1560,1561,1562,1563,1564,1565,1566,1567,1568,1569,1570,1571],{"id":1432,"depth":656,"text":1433},{"id":1439,"depth":656,"text":1440},{"id":1446,"depth":656,"text":1447},{"id":1455,"depth":656,"text":1456},{"id":1462,"depth":656,"text":1463},{"id":1471,"depth":656,"text":1472},{"id":1478,"depth":656,"text":1479},{"id":1485,"depth":656,"text":1486},{"id":1492,"depth":656,"text":1493},{"id":1499,"depth":656,"text":1500},{"id":1506,"depth":656,"text":1507},{"id":1513,"depth":656,"text":1514},{"id":607,"depth":156,"text":608},{"id":618,"depth":156,"text":619},"Institutional memory is what the company still knows after the person who did the work leaves: official playbooks, signed decisions, and live systems — with access control.","/blog/what-is-institutional-memory-in-enterprise-ai",{"title":1129,"description":1574},"blog/what-is-institutional-memory-in-enterprise-ai",[679,1579,1580,938],"institutional-memory","lifecycle-graph","QKskx6GjKfWSaN9E9kbMHU2XCu29nlgWTKO656jg8aI",{"enabled":149,"message":1583,"linkLabel":78,"linkHref":79,"id":1584,"title":1585,"archived":149,"authors":150,"badge":150,"body":1586,"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":1590,"relatedHeading":150,"seo":1591,"series":150,"sitemap":115,"status":150,"stem":1592,"subhead":150,"tags":150,"video":150,"whyJoin":150,"workplaceType":150,"__hash__":1593},"We're hiring! Join the team building the Sentient Enterprise.","content/shared/hiring.md","Hiring banner",{"type":152,"value":1587,"toc":1588},[],{"title":155,"searchDepth":156,"depth":156,"links":1589},[],"/shared/hiring",{"title":1585,"description":155},"shared/hiring","-6bioYD7lKYokGUVU3ff4hHTvB-sDyOMuCptKHnojfk",{"fold":1595,"id":1599,"title":1600,"archived":149,"authors":150,"badge":150,"body":1601,"date":150,"department":150,"description":155,"extension":158,"eyebrow":150,"faqHeader":150,"faqs":150,"footerBand":1605,"headline":150,"image":150,"industry":150,"jobType":150,"listed":149,"location":150,"navigation":115,"openRoles":150,"pageLayout":150,"path":1609,"relatedHeading":150,"seo":1610,"series":150,"sitemap":115,"status":150,"stem":1611,"subhead":150,"tags":150,"video":150,"whyJoin":150,"workplaceType":150,"__hash__":1612},{"headline":1596,"description":1597,"primaryLabel":8,"primaryTo":1598,"secondaryLabel":701,"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":1602,"toc":1603},[],{"title":155,"searchDepth":156,"depth":156,"links":1604},[],{"headline":1606,"description":1607,"primaryLabel":8,"primaryTo":1598,"secondaryLabel":1608,"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":1600,"description":155},"shared/cta","YHK6Fb8AvCPR1zZq7R_xiXUG0hwhP5UxHA8Ix52JQp4",1787194074224]