[{"data":1,"prerenderedAt":2058},["ShallowReactive",2],{"site-nav-content":3,"blog:\u002Fblog\u002Fwhat-is-an-agent-harness":179,"blog-index-copy":908,"blog:\u002Fblog\u002Fwhat-is-an-agent-harness:surround":929,"hiring-banner-content":2027,"site-cta-content":2039},{"header":4,"productNav":9,"nav":42,"footer":61,"askAI":132,"id":163,"title":164,"archived":165,"authors":166,"badge":166,"body":167,"date":166,"definedTerm":166,"department":166,"description":171,"extension":174,"eyebrow":166,"faqHeader":166,"faqs":166,"footerBand":166,"headline":166,"image":166,"industry":166,"jobType":166,"listed":131,"location":166,"navigation":131,"openRoles":166,"pageLayout":166,"path":175,"relatedHeading":166,"seo":176,"series":166,"sitemap":165,"status":166,"stem":177,"subhead":166,"tags":166,"video":166,"whyJoin":166,"workplaceType":166,"__hash__":178},{"productLabel":5,"loginLabel":6,"contactLabel":7,"contactSalesLabel":8},"Product","Log in","Contact","Get started for free",[10,14,18,22,26,30,34,38],{"label":11,"to":12,"description":13},"Overview","\u002Foverview","Seven layers. One closed loop.",{"label":15,"to":16,"description":17},"Conflux","\u002Fproduct\u002Fconflux","Where your team, workstreams, and agents meet.",{"label":19,"to":20,"description":21},"Agent Teams","\u002Fproduct\u002Fagent-teams","Specialist teams - governed from day one.",{"label":23,"to":24,"description":25},"Lifecycle Graph","\u002Fproduct\u002Flifecycle-graph","Intelligence that compounds across every interaction.",{"label":27,"to":28,"description":29},"Company Wiki","\u002Fproduct\u002Fwiki","Playbooks and policies where expertise stays.",{"label":31,"to":32,"description":33},"Workstreams","\u002Fproduct\u002Fworkstreams","From brief to signed-off deliverable on one canvas.",{"label":35,"to":36,"description":37},"Perception Console","\u002Fproduct\u002Fperception","Ask your whole business in plain English.",{"label":39,"to":40,"description":41},"Governance","\u002Fproduct\u002Fgovernance","Frontier AI you can actually sign off on.",[43,46,49,52,55,58],{"label":44,"to":45},"Models","\u002Fmodels",{"label":47,"to":48},"Pricing","\u002Fpricing",{"label":50,"to":51},"Integrations","\u002Fintegrations",{"label":53,"to":54},"Security","\u002Fsecurity",{"label":56,"to":57},"Partners","\u002Fpartners",{"label":59,"to":60},"Insights","\u002Fblog",{"productHeading":5,"companyHeading":62,"resourcesHeading":63,"legalHeading":64,"docsLabel":65,"docsUrl":66,"statementLines":67,"copyright":70,"companyLinks":71,"resourcesLinks":86,"legalLinks":102,"socialLinks":109,"bottomLinks":119},"Company","Resources","Legal","Docs","https:\u002F\u002Fdocs.gonimbus.ai",[68,69],"Stop training someone else's model.","Control your AI.","© 2026 Nimbus Intelligence, Inc. All rights reserved.",[72,73,74,75,76,78,81,84],{"label":47,"to":48},{"label":50,"to":51},{"label":53,"to":54},{"label":59,"to":60},{"label":77,"to":57},"Partner Program",{"label":79,"to":80},"Careers","\u002Fcareers",{"label":82,"to":83},"System status","\u002Fstatus",{"label":7,"to":85},"\u002Fcontact",[87,90,93,96,99],{"label":88,"to":89},"Glossary","\u002Fglossary",{"label":91,"to":92},"Compare","\u002Fcompare",{"label":94,"to":95},"Evaluate","\u002Fevaluate",{"label":97,"to":98},"Problems","\u002Fproblems",{"label":100,"to":101},"Use cases","\u002Fuse-cases",[103,106],{"label":104,"to":105},"Terms of Service","\u002Fterms",{"label":107,"to":108},"Privacy Policy","\u002Fprivacy",[110,113,116],{"label":111,"href":112},"LinkedIn","https:\u002F\u002Fwww.linkedin.com\u002Fcompany\u002Fgonimbusai\u002F",{"label":114,"href":115},"X","https:\u002F\u002Fx.com\u002Fgonimbusai",{"label":117,"href":118},"Instagram","https:\u002F\u002Fwww.instagram.com\u002Fgonimbus_ai\u002F",[120,122,124,127,128],{"label":121,"to":105},"Terms",{"label":123,"to":108},"Privacy",{"label":125,"to":126},"Compliance","\u002Fcompliance",{"label":82,"to":83},{"label":129,"to":130,"external":131},"LLMs.txt","\u002Fllms.txt",true,{"text":133,"prompt":134},"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":135,"platforms":137},{" Summarize the highlights from Nimbus's website":136},"https:\u002F\u002Fgonimbus.ai",[138,143,148,153,158],{"name":139,"label":140,"icon":141,"hrefPrefix":142},"chatgpt","ChatGPT","simple-icons:openai","https:\u002F\u002Fchatgpt.com\u002F?prompt=",{"name":144,"label":145,"icon":146,"hrefPrefix":147},"perplexity","Perplexity","mdi:magnify","https:\u002F\u002Fwww.perplexity.ai\u002Fsearch\u002Fnew?q=",{"name":149,"label":150,"icon":151,"hrefPrefix":152},"grok","Grok","simple-icons:x","https:\u002F\u002Fx.com\u002Fi\u002Fgrok?text=",{"name":154,"label":155,"icon":156,"hrefPrefix":157},"claude","Claude","simple-icons:anthropic","https:\u002F\u002Fclaude.ai\u002Fnew?q=",{"name":159,"label":160,"icon":161,"hrefPrefix":162},"google-ai","Google AI","simple-icons:google","https:\u002F\u002Fwww.google.com\u002Fsearch?udm=50&aep=11&q=","content\u002Fshared\u002Fnav.md","Site navigation",false,null,{"type":168,"value":169,"toc":170},"minimark",[],{"title":171,"searchDepth":172,"depth":172,"links":173},"",2,[],"md","\u002Fshared\u002Fnav",{"title":164,"description":171},"shared\u002Fnav","1dD7ahDRl0SQ4hz53-kKo0tEFrGaLuaztZ3PPfp6a9k",{"id":180,"title":181,"archived":165,"authors":182,"badge":185,"body":187,"date":897,"definedTerm":166,"department":166,"description":898,"extension":174,"eyebrow":166,"faqHeader":166,"faqs":166,"footerBand":166,"headline":166,"image":166,"industry":166,"jobType":166,"listed":131,"location":166,"navigation":131,"openRoles":166,"pageLayout":166,"path":899,"relatedHeading":166,"seo":900,"series":901,"sitemap":131,"status":166,"stem":902,"subhead":166,"tags":903,"video":166,"whyJoin":166,"workplaceType":166,"__hash__":907},"content\u002Fblog\u002Fwhat-is-an-agent-harness.md","What is an Agent Harness",[183],{"name":184,"to":136},"Nimbus Research",{"label":186},"Explainer",{"type":168,"value":188,"toc":878},[189,210,219,233,251,256,369,386,390,398,401,418,433,452,460,464,467,483,495,513,526,536,552,567,571,574,577,585,593,613,635,639,651,660,671,678,682,687,690,694,697,701,709,713,716,720,727,731,738,742,755,759,771,775],[190,191,192,193,197,198,205,206,209],"p",{},"An ",[194,195,196],"strong",{},"agent harness"," is the software around a large language model that turns next-token prediction into work: tools, memory, a loop, permissions, and a stop. ",[199,200,204],"a",{"href":201,"rel":202},"https:\u002F\u002Fdocs.langchain.com\u002Foss\u002Fpython\u002Flangchain\u002Fagents",[203],"nofollow","LangChain’s 2026 documentation"," writes the equation in plain type: ",[194,207,208],{},"Agent = Model + Harness",". The model reasons. The harness is everything else.",[190,211,212,213,218],{},"That sentence is not marketing. An unaided model is stateless. It produces text. It cannot keep a file, call Salesforce, fail a linter, or refuse a write. The ",[199,214,217],{"href":215,"rel":216},"https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FAgent_harness",[203],"Wikipedia entry on agent harnesses"," records the same split, and notes that the UK’s AI Security Institute already described an AI agenthe model plus scaffolding in 2023. The industry spent two years arguing about which model was smartest. In 2026 it started arguing about which environment the model was sitting in.",[190,220,221,226,227,232],{},[199,222,225],{"href":223,"rel":224},"https:\u002F\u002Fwww.databricks.com\u002Fblog\u002Fai-harness",[203],"Databricks"," uses a body-and-brain analogy: the model is the brain; the harness is the body and the workspace. ",[199,228,231],{"href":229,"rel":230},"https:\u002F\u002Fwww.langchain.com\u002Fblog\u002Fthe-anatomy-of-an-agent-harness",[203],"LangChain’s anatomy post"," is more mechanical. A harness is every piece of code, configuration, and execution logic that is not the model itself. A raw model is not an agent. It becomes one when a harness gives it state, tool execution, feedback loops, and constraints that do not depend on the model’s mood.",[190,234,235,236,240,241,245,246,250],{},"This article is the definition. ",[199,237,239],{"href":238},"what-is-harness-engineering","What is harness engineering"," is the practice of tightening that environment when the agent fails. ",[199,242,244],{"href":243},"inner-vs-outer-agent-harness","Inner vs outer agent harness"," is the cut between a repo and a company. An ",[199,247,249],{"href":248},"what-is-an-enterprise-agent-harness","enterprise agent harness"," is the outer case: operators, signers, a ledger.",[252,253,255],"h2",{"id":254},"words-youll-hear","Words you’ll hear",[257,258,259,266,272,278,290,307,328,344,353],"ul",{},[260,261,262,265],"li",{},[194,263,264],{},"Harness \u002F scaffolding."," Same object, two eras. Scaffolding is the 2023–2024 research word. Harness is the 2026 product word. Both mean the runtime around the weights.",[260,267,268,271],{},[194,269,270],{},"Agent."," The composed system. Not the model. Not the chat UI. If you can swap the model and the job still runs, you were looking at the harness.",[260,273,274,277],{},[194,275,276],{},"Loop."," Plan, act, observe, repeat — the ReAct-shaped cycle popularised in 2022 and now owned by the harness, not by the prompt. The harness dispatches the tool, returns the result, and decides whether to continue.",[260,279,280,283,284,289],{},[194,281,282],{},"Stop."," Budget, max steps, empty retrieval, tool error, human cancel, wait-for-named-signer. “The model says it is done” is a suggestion. ",[199,285,288],{"href":286,"rel":287},"https:\u002F\u002Fwww.anthropic.com\u002Fengineering\u002Fbuilding-effective-agents",[203],"Anthropic’s note on building effective agents"," is honest about this: encoding the job and deciding what “done” means is the boring part that actually matters.",[260,291,292,295,296,301,302,306],{},[194,293,294],{},"Tools \u002F skills \u002F MCP."," Hands. The ",[199,297,300],{"href":298,"rel":299},"https:\u002F\u002Fmodelcontextprotocol.io\u002Fspecification\u002F2025-11-25\u002Findex",[203],"Model Context Protocol"," is a common plug so hosts can call the same servers. Plumbing. A plug is not a permission model. See ",[199,303,305],{"href":304},"what-is-model-context-protocol","What is Model Context Protocol",".",[260,308,309,312,313,318,319,323,324,327],{},[194,310,311],{},"Hooks \u002F middleware."," Deterministic intercepts on the loop. ",[199,314,317],{"href":315,"rel":316},"https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fhooks",[203],"Claude Code hooks"," run shell or HTTP at ",[320,321,322],"code",{},"PreToolUse"," and can block with exit code 2. LangChain middleware is the same instinct in a library. A line in ",[320,325,326],{},"CLAUDE.md"," is advice. A hook is a gate.",[260,329,330,333,334,339,340,343],{},[194,331,332],{},"Guides and sensors."," ",[199,335,338],{"href":336,"rel":337},"https:\u002F\u002Fmartinfowler.com\u002Farticles\u002Fharness-engineering.html",[203],"Birgitta Böckeler’s framing on martinfowler.com",": feed-forward context (conventions, architecture, ",[320,341,342],{},"AGENTS.md",") versus feedback (linters, tests, reviewers). A harness that only prompts is half a harness.",[260,345,346,349,350,306],{},[194,347,348],{},"Inner harness."," Coding agents: Claude Code, Cursor, Codex. Workspace is a repository. Tests are the eval. See ",[199,351,352],{"href":243},"inner vs outer",[260,354,355,358,359,363,364,368],{},[194,356,357],{},"Outer \u002F enterprise harness."," Operators. Connectors, ",[199,360,362],{"href":361},"what-is-an-ai-workstream","workstreams",", ",[199,365,367],{"href":366},"what-is-write-back-governance","write-back",", a ledger. Workspace is the company. A passing unit test does not prove a CRM write was authorised.",[190,370,371,372,363,375,363,378,363,380,363,383,385],{},"Nimbus is one outer harness: ",[199,373,374],{"href":28},"wiki",[199,376,377],{"href":20},"agent teams",[199,379,362],{"href":32},[199,381,382],{"href":40},"governance",[199,384,23],{"href":24},". Claude Code is a strong inner harness. Calling either “an agent” without naming the harness is how RFPs buy a model and inherit someone else’s loop.",[252,387,389],{"id":388},"why-you-should-care","Why you should care",[190,391,392,397],{},[199,393,396],{"href":394,"rel":395},"https:\u002F\u002Fwww.mckinsey.com\u002Fcapabilities\u002Fquantumblack\u002Four-insights\u002Fthe-state-of-ai",[203],"McKinsey’s 2025 State of AI survey"," is the scale gap in one chart: most organisations use AI in at least one function; far fewer have begun to scale. Copilots produce usage. Harnesses produce jobs that finish under a stop. If your programme is “we rolled out ChatGPT Enterprise,” you have licensed a model surface. You have not yet chosen a harness for the work that writes back.",[190,399,400],{},"It affects you if:",[257,402,403,406,409,412,415],{},[260,404,405],{},"the job is multi-step and tool-using, not a single completion",[260,407,408],{},"a live system can change (CRM, ERP, repo, ticket queue)",[260,410,411],{},"someone will ask, six months later, why a field or a file changed",[260,413,414],{},"you need to swap models without rewriting every tool",[260,416,417],{},"you already noticed that a better model still skips the linter, invents a policy, or pastes into Salesforce",[190,419,420,421,426,427,432],{},"In February 2024 a British Columbia tribunal held Air Canada responsible for a chatbot that invented a bereavement-fare policy. ",[199,422,425],{"href":423,"rel":424},"https:\u002F\u002Fwww.cbc.ca\u002Fnews\u002Fcanada\u002Fbritish-columbia\u002Fair-canada-chatbot-lawsuit-1.7116416",[203],"CBC reported"," that the airline’s argument — the chatbot is a separate legal entity — failed. That failure is a missing harness, not a missing model: no quote, no signer, no stop before a commitment left the building. In 2023 a New York court ",[199,428,431],{"href":429,"rel":430},"https:\u002F\u002Fwww.reuters.com\u002Flegal\u002Fnew-york-lawyers-sanctioned-using-fake-chatgpt-cases-legal-brief-2023-06-22\u002F",[203],"sanctioned lawyers"," who filed ChatGPT-invented cases. Ungated generation reached a system of record. CRM writes are the operational twin with money attached.",[190,434,435,440,441,446,447,451],{},[199,436,439],{"href":437,"rel":438},"https:\u002F\u002Fwww.nist.gov\u002Fitl\u002Fai-risk-management-framework",[203],"NIST’s AI RMF"," organises Govern, Map, Measure, Manage. ",[199,442,445],{"href":443,"rel":444},"https:\u002F\u002Fwww.iso.org\u002Fstandard\u002F42001",[203],"ISO\u002FIEC 42001"," is an AI ",[448,449,450],"em",{},"management system"," standard. Neither is implemented by a system prompt that says “be careful.” They are implemented by a runtime that can refuse a tool call.",[190,453,454,459],{},[199,455,458],{"href":456,"rel":457},"https:\u002F\u002Faddyosmani.com\u002Fblog\u002Fagent-harness-engineering\u002F",[203],"Addy Osmani’s 2026 write-up"," states the engineering claim operators keep rediscovering: a decent model with a great harness beats a great model with a bad harness. When the agent does something dumb, the default instinct is to blame the weights. Harness engineering treats most of those failures as configuration. That is the rest of this cluster.",[252,461,463],{"id":462},"what-a-harness-actually-contains","What a harness actually contains",[190,465,466],{},"LangChain’s anatomy and Databricks’s list converge on the same parts. You can inspect each one before you buy a product or assemble a library.",[190,468,469,472,473,478,479,482],{},[194,470,471],{},"The loop."," The harness owns plan → act → observe. It executes the tool. It feeds the result back. It enforces max steps and a cost budget so a stuck agent cannot run forever. ",[199,474,477],{"href":475,"rel":476},"https:\u002F\u002Fwww.anthropic.com\u002Fengineering\u002Feffective-harnesses-for-long-running-agents",[203],"Anthropic’s long-running harness note"," shows why this is not a prompt: tasks that outlast one context window need an initializer, incremental sessions, git commits, and a progress file the ",[448,480,481],{},"next"," session can read. The model does not remember. The environment does.",[190,484,485,488,489,494],{},[194,486,487],{},"Tools and execution."," Search, shell, apply_patch, browser, CRM, ERP. The model proposes a call. The harness runs it in a sandbox or against an adapter, handles timeouts, and returns structured results. A generic HTTP tool with a production token is not a harness. It is a confused deputy. ",[199,490,493],{"href":491,"rel":492},"https:\u002F\u002Fgenai.owasp.org\u002Fllm-top-10\u002F",[203],"OWASP’s Top 10 for LLM applications"," still applies: excessive agency and unbounded tool use are design failures, not model quirks.",[190,496,497,500,501,503,504,508,509,306],{},[194,498,499],{},"Context and memory."," Working memory is the current window. Session state is progress for this job. Durable memory is files, ",[320,502,342],{},", a wiki, or a graph — something that survives compaction. Anthropic’s initializer\u002Fcoding-agent split is a memory design: feature lists and commits as cross-session state. A company that stores “what we approved” only in Slack search does not have durable memory for operations. See ",[199,505,507],{"href":506},"what-is-a-lifecycle-graph","What is a lifecycle graph"," and ",[199,510,512],{"href":511},"what-is-a-company-wiki-for-ai-agents","What is a company wiki for AI agents",[190,514,515,518,519,521,522,525],{},[194,516,517],{},"Permissions and hooks."," Who may call which tool, with which identity, on which object. Claude Code’s ",[320,520,322],{}," hook can deny Bash regardless of what the model intended. That is the inner version of ",[199,523,524],{"href":366},"write-back governance",": the write API is unreachable until a named role signs a quoted payload. A prompt that says “ask Legal first” is not this layer. The model can forget. The user can paste anyway.",[190,527,528,531,532,306],{},[194,529,530],{},"Feedback."," Compilers, tests, linters, schema validators, human review. Böckeler’s sensors. Without them the loop is open: the model reports success and the harness believes it. Terminal-Bench and SWE-bench exist because coding harnesses can grade against an environment. Enterprise writes need an equivalent: did the signed payload match what executed. See ",[199,533,535],{"href":534},"eval-loops-for-enterprise-agent-harnesses","eval loops for enterprise agent harnesses",[190,537,538,541,542,546,547,551],{},[194,539,540],{},"Orchestration."," Subagents, hand-offs, model routing. Optional until the job already splits in the organisation. ",[199,543,545],{"href":544},"what-is-multi-agent-ai","Multi-agent AI"," is the pattern. ",[199,548,550],{"href":549},"agent-team-architecture","Agent team architecture"," is the hiring object. A harness that spawns specialists without a stop is a faster way to share a production login.",[190,553,192,554,558,559,562,563,566],{},[199,555,557],{"href":556},"what-is-an-agentic-workflow","agentic workflow"," is a designed sequence with business stops. The harness is the runtime that can actually run that sequence. Mixing those two words is how demos skip isolation. A ",[199,560,561],{"href":361},"workstream"," is the company object that hosts the job: brief, connectors, people, budget. In Nimbus the workstream is that folder; the harness is wiki + teams + connectors + gates + graph sitting around whichever model ",[199,564,565],{"href":45},"routing"," picks for the step.",[252,568,570],{"id":569},"what-is-not-a-harness","What is not a harness",[190,572,573],{},"A chat window with plugins. The human is still the message bus, the permission system, and the audit log.",[190,575,576],{},"A system prompt. Advice inside the window. Useful. Not a stop.",[190,578,579,580,584],{},"A policy PDF. ",[199,581,583],{"href":582},"what-is-ai-governance","What is AI governance"," is a management claim. A harness is whether an unapproved write is impossible.",[190,586,587,588,592],{},"MCP on its own. A standard plug. See ",[199,589,591],{"href":590},"mcp-for-enterprise-integrations","MCP for enterprise integrations",". If the server can PATCH Salesforce from natural language, you built a bypass.",[190,594,595,596,604,605,608,609,306],{},"A framework on its own. ",[199,597,600,601],{"href":598,"rel":599},"https:\u002F\u002Fwww.langchain.com\u002Fblog\u002Fhow-to-build-a-custom-agent-harness",[203],"LangChain’s ",[320,602,603],{},"create_agent"," is a way to ",[448,606,607],{},"assemble"," a harness. CrewAI, LangGraph, and Pydantic AI are in the same neighbourhood. You still have to choose tools, stops, and identity. See ",[199,610,612],{"href":611},"agent-harness-vs-agent-framework","agent harness vs agent framework",[190,614,615,616,508,621,626,627,508,631,306],{},"A copilot seat. ",[199,617,620],{"href":618,"rel":619},"https:\u002F\u002Fopenai.com\u002Fbusiness\u002Fchatgpt-enterprise\u002F",[203],"ChatGPT Enterprise",[199,622,625],{"href":623,"rel":624},"https:\u002F\u002Fwww.microsoft.com\u002Fen-us\u002Fmicrosoft-365\u002Fcopilot",[203],"Microsoft 365 Copilot"," are excellent personal surfaces. They are not, by default, a company loop with fail-closed writes. See ",[199,628,630],{"href":629},"how-to-choose-between-a-copilot-and-a-work-os","How to choose between a copilot and a work OS",[199,632,634],{"href":633},"how-to-choose-between-a-coding-harness-and-an-enterprise-harness","How to choose between a coding harness and an enterprise harness",[252,636,638],{"id":637},"how-this-shows-up-in-products","How this shows up in products",[190,640,641,644,645,647,648,650],{},[194,642,643],{},"Coding harnesses."," Claude Code, Cursor, Codex, open shells like OpenHands. Workspace is a checkout. ",[320,646,326],{}," \u002F ",[320,649,342],{}," are guides. Hooks, tests, and CI are sensors. Eval is SWE-bench or Terminal-Bench, or your own suite. These are the right shape for software.",[190,652,653,656,657,659],{},[194,654,655],{},"Library harnesses."," LangChain ",[320,658,603],{},", Deep Agents, LangGraph graphs. You compose the loop in code. You own production identity. Good when the job is yours to engineer. A liability when operators are expected to “just add Salesforce.”",[190,661,662,665,666,670],{},[194,663,664],{},"Enterprise \u002F outer harnesses."," Palantir AIP, Salesforce Agentforce, and self-service OS-class products such as Nimbus. The workspace is a job with connectors and people, not a git root. The interesting stop is a named signer on a quoted write, not a green test. ",[199,667,669],{"href":668},"how-to-evaluate-an-agent-harness","How to evaluate an agent harness"," is the buying sheet.",[190,672,673,674,677],{},"Nimbus’s mapping is deliberate and not unique as a ",[448,675,676],{},"category",": Perception orients, Conflux collaborates, agent teams run, governance quotes, the graph records. You can score that mapping against the parts above. You should score AIP and Agentforce the same way. Category names do not substitute for a failed write.",[252,679,681],{"id":680},"questions-people-actually-ask","Questions people actually ask",[683,684,686],"h3",{"id":685},"is-the-model-the-agent","Is the model the agent?",[190,688,689],{},"No. The agent is model plus harness. Shopping for a model is shopping for a chip. Shopping for a harness is shopping for how work finishes.",[683,691,693],{"id":692},"do-i-need-a-harness-for-a-single-prompt","Do I need a harness for a single prompt?",[190,695,696],{},"No. A completion does not need a loop. Multi-step tool use does. Long-running work that outlasts one window does. Writes to live systems do.",[683,698,700],{"id":699},"is-rag-a-harness","Is RAG a harness?",[190,702,703,704,708],{},"Retrieval is a tool and a memory pattern inside a step. ",[199,705,707],{"href":706},"what-is-enterprise-rag","Enterprise RAG"," does not dispatch tools, enforce a signer, or persist a decision. Useful. Incomplete.",[683,710,712],{"id":711},"can-i-just-use-mcp-as-my-harness","Can I just use MCP as my harness?",[190,714,715],{},"You can use MCP as the plug. You still need identity, scope, quoting, and a stop. The spec does not require those.",[683,717,719],{"id":718},"will-a-better-model-shrink-the-harness","Will a better model shrink the harness?",[190,721,722,726],{},[199,723,725],{"href":456,"rel":724},[203],"Osmani"," and Anthropic’s long-running work both say the ceiling moves. Tasks that were unreachable come into play and bring new failure modes. Stronger models still do not know your signer, your budget, or your CRM field map.",[683,728,730],{"id":729},"how-is-this-different-from-an-enterprise-ai-os","How is this different from an enterprise AI OS?",[190,732,192,733,737],{},[199,734,736],{"href":735},"what-is-an-enterprise-ai-operating-system","enterprise AI operating system"," is the company-shaped product: wiki, workstreams, teams, gates, ledger. A harness is the runtime idea underneath — including coding harnesses that are not an OS. Nimbus is an OS-class outer harness. Claude Code is not an OS. Both are harnesses.",[683,739,741],{"id":740},"what-should-i-read-next","What should I read next?",[190,743,744,746,747,751,752,754],{},[199,745,239],{"href":238}," for the practice. ",[199,748,750],{"href":749},"agent-harness-architecture","Agent harness architecture"," for the parts in one diagram. ",[199,753,669],{"href":668}," before a vendor demo.",[252,756,758],{"id":757},"related-reading","Related reading",[190,760,761,363,764,767,768,306],{},[199,762,763],{"href":556},"What is an agentic workflow",[199,765,766],{"href":544},"What is multi-agent AI",", and ",[199,769,770],{"href":366},"What is write-back governance",[252,772,774],{"id":773},"sources","Sources",[257,776,777,783,789,795,801,807,813,819,825,831,837,843,848,854,860,866,872],{},[260,778,779],{},[199,780,782],{"href":201,"rel":781},[203],"LangChain, Agents (Agent = Model + Harness)",[260,784,785],{},[199,786,788],{"href":229,"rel":787},[203],"LangChain, The anatomy of an agent harness",[260,790,791],{},[199,792,794],{"href":598,"rel":793},[203],"LangChain, How to build a custom agent harness",[260,796,797],{},[199,798,800],{"href":215,"rel":799},[203],"Wikipedia, Agent harness",[260,802,803],{},[199,804,806],{"href":223,"rel":805},[203],"Databricks, What is an AI agent harness?",[260,808,809],{},[199,810,812],{"href":286,"rel":811},[203],"Anthropic, Building effective agents",[260,814,815],{},[199,816,818],{"href":475,"rel":817},[203],"Anthropic, Effective harnesses for long-running agents",[260,820,821],{},[199,822,824],{"href":336,"rel":823},[203],"Böckeler, Harness engineering for coding agent users",[260,826,827],{},[199,828,830],{"href":456,"rel":829},[203],"Addy Osmani, Agent harness engineering",[260,832,833],{},[199,834,836],{"href":394,"rel":835},[203],"McKinsey, The state of AI in 2025",[260,838,839],{},[199,840,842],{"href":437,"rel":841},[203],"NIST AI Risk Management Framework",[260,844,845],{},[199,846,445],{"href":443,"rel":847},[203],[260,849,850],{},[199,851,853],{"href":423,"rel":852},[203],"CBC, Air Canada chatbot lawsuit",[260,855,856],{},[199,857,859],{"href":429,"rel":858},[203],"Reuters, New York lawyers sanctioned over ChatGPT citations",[260,861,862],{},[199,863,865],{"href":491,"rel":864},[203],"OWASP Top 10 for LLM applications",[260,867,868],{},[199,869,871],{"href":298,"rel":870},[203],"Model Context Protocol specification (2025-11-25)",[260,873,874],{},[199,875,877],{"href":315,"rel":876},[203],"Claude Code, Hooks",{"title":171,"searchDepth":172,"depth":172,"links":879},[880,881,882,883,884,885,895,896],{"id":254,"depth":172,"text":255},{"id":388,"depth":172,"text":389},{"id":462,"depth":172,"text":463},{"id":569,"depth":172,"text":570},{"id":637,"depth":172,"text":638},{"id":680,"depth":172,"text":681,"children":886},[887,889,890,891,892,893,894],{"id":685,"depth":888,"text":686},3,{"id":692,"depth":888,"text":693},{"id":699,"depth":888,"text":700},{"id":711,"depth":888,"text":712},{"id":718,"depth":888,"text":719},{"id":729,"depth":888,"text":730},{"id":740,"depth":888,"text":741},{"id":757,"depth":172,"text":758},{"id":773,"depth":172,"text":774},"2026-08-24","An agent harness is everything around a model that lets it do work: tools, memory, permissions, loops, and stops — Agent = Model + Harness, not a chat window with plugins.","\u002Fblog\u002Fwhat-is-an-agent-harness",{"title":181,"description":898},"explainer","blog\u002Fwhat-is-an-agent-harness",[901,904,905,906],"agent-harness","agents","harness-engineering","3CDvdR09AOUKC2e_n9BjOHZfpkkyjMqsAzkONc8Lf0E",{"hero":909,"id":911,"title":912,"archived":165,"authors":166,"badge":166,"body":913,"date":166,"definedTerm":166,"department":166,"description":917,"extension":174,"eyebrow":918,"faqHeader":166,"faqs":166,"footerBand":919,"headline":166,"image":166,"industry":166,"jobType":166,"listed":131,"location":166,"navigation":131,"openRoles":166,"pageLayout":166,"path":60,"relatedHeading":925,"seo":926,"series":166,"sitemap":131,"status":166,"stem":927,"subhead":166,"tags":166,"video":166,"whyJoin":166,"workplaceType":166,"__hash__":928},{"filename":910},"u2221455217_Flat_design_of_a_futuristic_minimalist_landscape__5d589295-cdea-4ea9-a262-be766881accf_1.png","content\u002Fblog\u002Findex.md","Exploring the future of intelligence.",{"type":168,"value":914,"toc":915},[],{"title":171,"searchDepth":172,"depth":172,"links":916},[],"Deep dives into pre-cognitive intelligence, sentient enterprises, and the evolving landscape of AI-driven business transformation.","Latest Research",{"headline":920,"description":921,"primaryLabel":922,"primaryTo":923,"secondaryLabel":924,"secondaryTo":12},"Stay at the frontier.","Subscribe for product updates and new insights.","Subscribe","\u002Fnewsletter","Explore the platform","More research",{"title":912,"description":917},"blog\u002Findex","BFSWGYO9bcTlaulivKYWyg08_DJHsdGg3OC6g_CG1Hw",[930,1143],{"id":931,"title":932,"archived":165,"authors":933,"badge":935,"body":937,"date":1130,"definedTerm":166,"department":166,"description":1131,"extension":174,"eyebrow":166,"faqHeader":166,"faqs":166,"footerBand":166,"headline":166,"image":1132,"industry":166,"jobType":166,"listed":131,"location":166,"navigation":131,"openRoles":166,"pageLayout":166,"path":1133,"relatedHeading":166,"seo":1134,"series":166,"sitemap":131,"status":166,"stem":1135,"subhead":166,"tags":1136,"video":1141,"whyJoin":166,"workplaceType":166,"__hash__":1142},"content\u002Fblog\u002Frebuilding-trust-in-global-agri-food-supply-chains.md","Rebuilding Trust in Global Agri-Food Supply Chains",[934],{"name":184,"to":136},{"label":936},"Supply Chain",{"type":168,"value":938,"toc":1122},[939,942,945,949,952,955,959,962,965,969,972,975,978,982,985,988,991,995,998,1001,1004,1007,1011,1014,1019],[190,940,941],{},"Global agricultural supply chains today suffer from chronic mistrust and fragility. Climate change, extreme weather and geopolitical conflicts have generated unprecedented volatility in crop yields and prices. For example, droughts and floods in 2024 drove cereal yields far below historical averages in Africa and Europe, and sharp weather-induced shortages sent cocoa prices surging 400%. Meanwhile, pandemic lockdowns and logistics failures disrupted labour, processing and transport on a massive scale. These shocks reveal the interdependence of farmers, traders, manufacturers and retailers around the world, and also how opacity and disorganization have allowed even small crises to ripple into full-blown system shocks. As one industry commentator noted, such events are “stress signals from a global system stretched beyond resilience”. The food sector still lacks real-time visibility into how ingredients move through thousands of suppliers, and data remain “fragmented across thousands of suppliers and opaque standards”. In this environment of uncertainty, stakeholders cannot easily verify risks or coordinate responses, so trust among partners has eroded.",[190,943,944],{},"For finance chiefs, this trust deficit is particularly problematic. In recent years, CFOs have become de facto risk managers for enterprise resilience, accountable not just for budgets but for the continuity of global supply lines. The disruptions of COVID-19, trade wars and climate shocks have shown CFOs that “confidence in suppliers’ capability, reliability, … and transparency becomes critical”. CFOs increasingly must ensure that suppliers have robust contingency plans and that potential exposures are identified early to address enterprise risk. Indeed, Deloitte reports that trust-building investments correlate with far greater supply-chain resiliency and even significant revenue growth. Yet many leaders also admit to blind spots: one survey found executives overestimate the trustworthiness of their chains by 20% on average. With such high stakes - and with extreme events becoming more frequent - today’s CFO must take a leading role in diagnosing and mitigating long-range supply chain risk.",[252,946,948],{"id":947},"the-evolving-cfo-mandate-in-supply-chain-resilience","The Evolving CFO Mandate in Supply Chain Resilience",[190,950,951],{},"Traditionally, CFOs focused on short-term financial performance. In a volatile post-pandemic era, however, the remit of the CFO has expanded into strategic risk and operations. Modern CFOs are “operating at the center of disruption - managing economic volatility, shifting trade and tax policy, and rapid advances in AI and emerging technology”. They are expected to align capital allocation with enterprise strategy, balancing investments in growth versus resilience. In practice, this means funding innovation in data systems, scenario planning, and cross-functional planning tools. It also means tightening financial discipline while supporting new business models and compliance demands. Our connected world means one misstep in supply procurement can be a multi-million-dollar problem, so CFOs must now account for long-tail supply risks in forecasts, disclosures, and budgeting. For example, 58% of surveyed CFOs say they are putting more emphasis on cash and liquidity forecasting to adjust to today’s volatility.",[190,953,954],{},"The reason is clear: supply chains are a major risk to business value. As Deloitte advises, CFOs should not assume their suppliers will simply weather crises on their own – instead, finance leaders should demand “well-designed, consistent plans” across the network to protect the firm against shocks. This means coordinating with procurement, operations and even external partners. CFOs who embrace transparency can better fulfill their mandate of enterprise risk oversight: by uncovering hidden exposures early, they can guide capital to the most resilient parts of the chain. On the other hand, CFOs who lack insight into supply linkages may overlook embedded risks. Indeed, finance chiefs who champion data-driven visibility enable faster, more informed decisions in turbulent times.",[252,956,958],{"id":957},"simulation-driven-decision-making-digital-twins-and-scenario-planning","Simulation-Driven Decision-Making: Digital Twins and Scenario Planning",[190,960,961],{},"To bridge the information gaps plaguing agri-food systems, many companies are adopting simulation platforms – essentially digital replicas of real-world supply networks and processes. At the core of this approach is the “digital twin” concept: a dynamic, data-driven model that mirrors physical assets, from farm equipment and silos to transport fleets and retail outlets. These virtual twins integrate real-time IoT sensor data, historical records and external feeds (weather, market indices, etc.) to represent the current state of the chain, and then run predictive models for the future. In agriculture, scholars note that digital twins can capture agronomic details like irrigation or fertilizer use and simulate crop growth and yield outcomes. By encompassing post-harvest steps – warehousing, distribution, processing – these systems can optimise the entire supply chain end-to-end.",[190,963,964],{},"Connected simulation is the next step: CFOs and planners feed these digital twins with proposed changes or disruptions (for example, a sudden trade embargo or a predicted drought) and see the virtual consequences. This scenario planning makes it possible to run “what-if” analyses that were previously impossible to manage manually. For instance, recent research highlights how a financial digital twin can combine operational and market data so that companies can simulate how, say, interest-rate swings or port closures would impact cash flows and working capital. In practical terms, digital twins allow companies to build rich “risk maps” of their multi-tier supplier networks and then stress-test them under various shocks. The technology thus provides unprecedented visibility: companies can track each node and link in real time, instantly spotting bottlenecks or quality issues. In Exiger’s words, digital twins offer a “comprehensive and real-time view of the entire ecosystem, enabling precise decision-making, better risk mitigation and long-term business continuity”.",[252,966,968],{"id":967},"benefits-for-visibility-alignment-and-coordination","Benefits for Visibility, Alignment, and Coordination",[190,970,971],{},"Simulation-driven platforms create a shared intelligence across stakeholders. Instead of each division or partner having its own isolated numbers, everyone looks at the same virtual model. This alignment greatly enhances trust. For example, a digital supply-chain twin can “provide unprecedented visibility” into supplier performance, inventory status, and material flows. When issues arise – say, a supplier is hit by flooding – the system immediately flags the affected nodes. Operations and finance can then jointly evaluate options: Could we reroute shipments? Ramp up alternative sources? How would each choice affect cost, revenue, and service levels? By simulating these scenarios, managers turn abstract risks into quantified outcomes. A case in point is Walmart’s use of a digital supply-chain replica: by running simulated scenarios of varying demand or port outages, the company could gauge the effect on inventory and service, helping it fine-tune stocking and routing strategies.",[190,973,974],{},"Importantly, simulation platforms foster proactivity. Rather than reacting when a crisis hits, organizations can test contingency plans in advance. They can answer questions like: “If we lose 30% of crop volume due to heat stress, will our pricing buffer or our logistics redundancy be enough?” This capability builds confidence. One study notes that companies using such what-if models can “evaluate the effects of demand fluctuations, seasonal changes, or supply chain interruptions” before they occur. Another analysis emphasizes that these systems detect anomalies or patterns (e.g. gradually declining supplier performance) that would otherwise go unnoticed. In practice, teams using digital twins for scenario analysis move from “reaction to pre-approved playbooks tied to quantified outcomes”. In short, shared simulations make hidden risks visible and help executives coordinate faster. As Rule Ltd. observes, proactive risk mapping plus “scenario planning change the conversation, you see the network clearly, you simulate credible what-ifs, and you choose the lowest-regret path with finance and operations aligned”.",[190,976,977],{},"This alignment extends trust. When a CFO and an operations leader look at the same simulation output, they build consensus on the best plan. Rule Ltd. notes that digital twin–supported scenario models “turn debate into numbers your CFO and COO can approve,” and in turn build confidence among stakeholders. The result is fewer surprises and a clearer audit trail – in fact, companies report that employing these tools leads to “fewer surprises for the board and key customers”. By replacing manual guesswork with data-driven clarity, simulation platforms can thus repair fractured trust.",[252,979,981],{"id":980},"lessons-from-recent-disruptions","Lessons from Recent Disruptions",[190,983,984],{},"Numerous recent events underscore the need for this approach. The COVID-19 pandemic exemplified how a lack of shared intelligence can fragment trust. As OECD analysts have documented, lockdowns imposed “unprecedented stresses on food supply chains” – from labor shortages in fields to processing-plant shutdowns and cross-border logjams. In many countries grocery shelves briefly emptied not from shortages of food per se, but from disruptions in logistics and coordination. During those tense weeks, buyers and suppliers struggled on siloed forecasts and outdated charts. By the time detailed data trickled through, panic orders had been placed or cancelled, eroding relationships. Transparency deficits even forced farmers in some regions to dump milk or waste perishable crops because they could not reach markets, weakening trust between agricultural producers and processors.",[190,986,987],{},"Environmental shocks further illustrate the point. In 2022, for example, simultaneous droughts and conflicts in major grain regions around the world caused a sudden 110% jump in wheat prices. No single country could have anticipated this alone, but global market data revealed the combined threat. Yet many local buyers found themselves scrambling, unsure of how to allocate inventory or hedge costs. If they had had a shared simulation of supply and demand flows, they might have mitigated the scare. Likewise, when the Suez Canal briefly blocked trade, manufacturers that could overlay that risk on their supply chain models with alternative routes avoided lengthy shutdowns. Without a common platform for such intelligence, suppliers can experience false alarms and buyers can accuse sellers of “unreliability,” further corroding trust.",[190,989,990],{},"Commodity price volatility is another case. We have seen agricultural inputs spike wildly – cocoa prices went up 400% after storms, a top processor called it “unprecedented disruption”, and coffee jumped 40% in a year. These swings reflect complex, interwoven factors. Yet if downstream companies had continuously updated scenario models of climate impact and trade trends, they could share projections with farmers and financiers in real time. Instead, price shocks today often trigger finger-pointing (e.g. is the trader at fault, or the grower, or the speculator?). Shared simulation data would at least ensure that everyone is looking at the same demand curves and weather forecasts. As one industry report starkly put it, “visibility becomes power” when a crisis is systemic. Failure to share that visibility cedes power to speculation and rumor – the very opposite of trust.",[252,992,994],{"id":993},"the-cfo-as-champion-of-simulation-and-transparency","The CFO as Champion of Simulation and Transparency",[190,996,997],{},"In all these contexts, the CFO is uniquely positioned to champion simulation technologies and rebuild trust. As the finance executive responsible for planning and investor communication, the CFO can drive investment in the necessary digital platforms. By allocating capital to build or procure digital twins and scenario tools, the CFO commits the organization to transparency. For example, CFOs can ensure that integrated business planning (IBP) processes connect FP&A with operations, so that scenario outcomes flow into forecasts and budgets. They can demand that supply-chain data be integrated with finance systems (as the WSC conference paper suggests, to automatically sync inventories and payables in a unified model).",[190,999,1000],{},"Most importantly, CFOs can use these tools to transform risk disclosure and stakeholder engagement. Instead of simply reporting static risk factors in footnotes, a CFO might present quantified scenarios – “what if” analyses of crop failure or tariff changes – grounded in the shared digital model. This level of open forecasting builds credibility with regulators, lenders and investors, because it shows a concrete plan rather than vague assurances. Internally, it also builds trust with other departments: the CFO is effectively saying “here is how I see the chain, let us plan together,” which encourages others to share data and cooperate.",[190,1002,1003],{},"Global companies are already piloting such approaches. A recent Cognizant analysis notes that businesses integrating digital twins “empower organizations to design, monitor, analyze and optimize assets and operations in real time, resulting in more accurate decisions and more efficient operations”. In practice, a food manufacturer might simulate factory outputs under different power-shutdown scenarios, enabling the CFO to decide whether to invest in backup generators or insurance. A grain trader might digitalize its entire procurement network and simulate futures-market variations, helping the CFO align hedging strategies with supply routes. Perhaps the most vivid example comes from UNICEF’s work: by using real-time shared data for vaccine distribution, UNICEF’s supply chain team (with support from finance planners) was able to “predict, respond and maintain resilient supply networks” during a crisis. The key was open data exchange and strong governance – exactly the principles CFOs should embed in agricultural chains.",[190,1005,1006],{},"Looking ahead, CFOs should ensure that digital twin investments also serve broader sustainability and regulatory goals. Traceability systems (often backed by blockchain or knowledge-graph technology) can become part of the simulation framework, linking financial metrics to environmental or social data. For instance, a food retailer may digitally map carbon footprints of its suppliers; running scenarios can then show how changing sources might reduce emissions while affecting cost. This kind of joint financial-operational modeling supports ESG disclosure, further enhancing stakeholder trust.",[252,1008,1010],{"id":1009},"conclusion","Conclusion",[190,1012,1013],{},"Rebuilding trust in the global agri-food system will not happen through goodwill alone; it requires hard data and shared perspective. Simulation-driven decision-making offers exactly that: a single source of truth for complex, uncertain environments. By championing digital twins and scenario planning, CFOs can turn opacity into transparency. They can quantify risk, allocate capital to where it most strengthens resilience, and communicate with confidence. In doing so, they restore the confidence of suppliers, buyers, investors and regulators. As one advisory firm notes, investing in these trust-building technologies is linked to stronger resilience and even higher revenue. In today’s volatile world, CFOs who embrace simulation are not just safeguarding operations – they are investing in credibility, earning stakeholder trust one model run at a time.",[190,1015,1016],{},[194,1017,1018],{},"References:",[257,1020,1021,1037,1044,1055,1062,1069,1076,1083,1090,1097,1104,1115],{},[260,1022,1023,1028,1033,306],{},[199,1024,1027],{"href":1025,"rel":1026},"https:\u002F\u002Fwww.deloitte.com\u002Fus\u002Fen\u002Fprograms\u002Fchief-financial-officer\u002Farticles\u002Ffor-cfos-enhancing-supply-chain-performance-may-be-a-matter-of-trust.html",[203],"Deloitte Insights, “For CFOs, enhancing supply chain performance may be a matter of trust” (2023)",[199,1029,1032],{"href":1030,"rel":1031},"https:\u002F\u002Fwww.deloitte.com\u002Fus\u002Fen\u002Fprograms\u002Fchief-financial-officer\u002Farticles\u002Ffor-cfos-enhancing-supply-chain-performance-may-be-a-matter-of-trust.html#:~:text=Still%2C%20finance%20leaders%20may%20have,%C2%B9",[203],"deloitte.com",[199,1034,1032],{"href":1035,"rel":1036},"https:\u002F\u002Fwww.deloitte.com\u002Fus\u002Fen\u002Fprograms\u002Fchief-financial-officer\u002Farticles\u002Ffor-cfos-enhancing-supply-chain-performance-may-be-a-matter-of-trust.html#:~:text=Furthermore%2C%20a%C2%A0Deloitte%20Global%20survey%C2%A0found%20that,risks%20that%20can%20influence%20performance",[203],[260,1038,1039],{},[199,1040,1043],{"href":1041,"rel":1042},"https:\u002F\u002Fwww.pwc.com\u002Fus\u002Fen\u002Fexecutive-leadership-hub\u002Fcfo.html#:~:text=%3E%20%5B58,planning%20in%20today%E2%80%99s%20volatile%20environment",[203],"PwC, “What’s important to the CFO in 2026” (PwC CFO Agenda)",[260,1045,1046],{},[199,1047,1050,1051,1054],{"href":1048,"rel":1049},"https:\u002F\u002Fpmc.ncbi.nlm.nih.gov\u002Farticles\u002FPMC11100011\u002F#:~:text=Digital%20Twins%20have%20emerged%20as,of%20agricultural%20lifecycle%2C%20edaphic%2C%20phytotechnologic",[203],"Escriba et al., ",[448,1052,1053],{},"Digital Twins in Agriculture: Orchestration and Applications"," (ACS Sustainable Chem. Eng. 2024)",[260,1056,1057],{},[199,1058,1061],{"href":1059,"rel":1060},"https:\u002F\u002Fwww.exiger.com\u002Fperspectives\u002Funlocking-the-potential-of-supply-chain-digital-twins\u002F#:~:text=,allow%20proactive%20risk%20identification%20through",[203],"Exiger, “Unlocking the Potential of Supply Chain Digital Twins” (2024)",[260,1063,1064],{},[199,1065,1068],{"href":1066,"rel":1067},"https:\u002F\u002Finforms-sim.org\u002Fwsc24papers\u002Fcon335.pdf#:~:text=Envision%20a%20scenario%20where%20enterprises,Consider%20the",[203],"Guivant et al., “Financial Digital Twin in the Supply Chain” (Proc. Winter Simulation Conf. 2024)",[260,1070,1071],{},[199,1072,1075],{"href":1073,"rel":1074},"http:\u002F\u002Fruleltd.com",[203],"Rule Ltd., “Risk Mapping and Scenario Planning for Supply Chains” (ruleltd.com)",[260,1077,1078],{},[199,1079,1082],{"href":1080,"rel":1081},"https:\u002F\u002Fprism.sustainability-directory.com\u002Fscenario\u002Fdigital-twins-for-agricultural-supply-chain-resilience\u002F#:~:text=than%20any%20technologist%2C%20that%20the,blockade%20on%20a%20local%20market",[203],"Smith, “Digital Twins for Agricultural Supply Chain Resilience” (Sustainability Directory, Nov 2025)",[260,1084,1085],{},[199,1086,1089],{"href":1087,"rel":1088},"https:\u002F\u002Fplanet-a.medium.com\u002Fde-risking-the-food-supply-chain-faa54bde6f2f",[203],"“De-risking the food supply chain” (Planet A Ventures, Nov 2025)",[260,1091,1092],{},[199,1093,1096],{"href":1094,"rel":1095},"https:\u002F\u002Fwww.weforum.org\u002Fstories\u002F2025\u002F01\u002Fai-supply-chains\u002F#:~:text=In%202024%2C%20KPMG%20reported%20that,global%20supply%20chain%20infrastructure%20more",[203],"World Economic Forum, “AI will protect global supply chains from the next major shock” (Jan 2025)",[260,1098,1099],{},[199,1100,1103],{"href":1101,"rel":1102},"https:\u002F\u002Fh2020-demeter.eu\u002Ftrust-and-transparency-of-data-in-the-agri-food-supply-chain-with-origintrail\u002F#:~:text=OriginTrail%20Decentralized%20Knowledge%20Graph%20,can%20be%20used%20by%20various",[203],"Demeter Project (EU), “Trust and Transparency of Data in the agri-food supply chain” (OriginTrail blog)",[260,1105,1106],{},[199,1107,1110,1111,1114],{"href":1108,"rel":1109},"https:\u002F\u002Fwww.oecd.org\u002Fcontent\u002Fdam\u002Foecd\u002Fen\u002Fpublications\u002Freports\u002F2020\u002F06\u002Ffood-supply-chains-and-covid-19-impacts-and-policy-lessons_62c97266\u002F71b57aea-en.pdf#:~:text=The%20COVID,While%20the%20impacts%20of%20COVID%0219",[203],"OECD, ",[448,1112,1113],{},"Food Supply Chains and COVID-19: Impacts and Policy Lessons"," (2020)",[260,1116,1117],{},[199,1118,1121],{"href":1119,"rel":1120},"https:\u002F\u002Fwww.cognizant.com\u002Fnl\u002Fen\u002Finsights\u002Fblog\u002Farticles\u002Fharnessing-digital-twins-and-simulation-modelling-for-strategic-advantages#:~:text=Digital%20twin%20technology%20and%20simulation,dependencies%2C%20improve%20supply%20chain%20resilience",[203],"Cognizant (Benelux), “Harnessing digital twins and simulation modelling for strategic advantages” (Apr 2024)",{"title":171,"searchDepth":172,"depth":172,"links":1123},[1124,1125,1126,1127,1128,1129],{"id":947,"depth":172,"text":948},{"id":957,"depth":172,"text":958},{"id":967,"depth":172,"text":968},{"id":980,"depth":172,"text":981},{"id":993,"depth":172,"text":994},{"id":1009,"depth":172,"text":1010},"2025-12-10","How simulation-driven decision-making and digital twins can help rebuild trust and resilience in global agri-food supply chains.","\u002Fassets\u002Fimages\u002Fblog\u002Fu9471466259_a_tractor_spraying_pesticides_on_a_vegetable_field__a80f3491-5525-4c85-bf30-a30e6cd9f683.png","\u002Fblog\u002Frebuilding-trust-in-global-agri-food-supply-chains",{"title":932,"description":1131},"blog\u002Frebuilding-trust-in-global-agri-food-supply-chains",[1137,1138,1139,1140],"supply-chain","resilience","digital-twins","agri-food","https:\u002F\u002Fcdn.gonimbus.ai\u002Fassets\u002Fwebsite\u002Fvideo\u002Fu9471466259_a_tractor_spraying_pesticides_on_a_vegetable_field__a80f3491-5525-4c85-bf30-a30e6cd9f683u.mp4","fw6hSeVlQHYioZnqzNnbusj3GZ8TInN1GAjrvBxIOa4",{"id":1144,"title":1145,"archived":165,"authors":1146,"badge":1148,"body":1149,"date":2002,"definedTerm":2003,"department":166,"description":2004,"extension":174,"eyebrow":166,"faqHeader":2005,"faqs":2008,"footerBand":166,"headline":166,"image":166,"industry":166,"jobType":166,"listed":131,"location":166,"navigation":131,"openRoles":166,"pageLayout":166,"path":2018,"relatedHeading":166,"seo":2019,"series":901,"sitemap":131,"status":166,"stem":2020,"subhead":166,"tags":2021,"video":166,"whyJoin":166,"workplaceType":166,"__hash__":2026},"content\u002Fblog\u002Fwhat-is-loop-engineering.md","What is Loop Engineering",[1147],{"name":184,"to":136},{"label":186},{"type":168,"value":1150,"toc":1988},[1151,1157,1164,1167,1196,1215,1219,1230,1236,1250,1266,1287,1337,1349,1351,1404,1407,1409,1412,1418,1424,1435,1437,1464,1472,1478,1484,1488,1499,1509,1519,1529,1540,1543,1547,1553,1559,1569,1575,1588,1594,1604,1607,1611,1614,1620,1625,1630,1640,1655,1662,1671,1674,1678,1684,1690,1696,1704,1713,1719,1727,1735,1738,1742,1745,1753,1767,1778,1781,1785,1792,1842,1848,1851,1855,1868,1877,1892,1894,1920,1922,1984],[190,1152,1153,1156],{},[194,1154,1155],{},"Loop engineering"," is the discipline of compiling the known path for unattended work, detecting when reality diverges from that path, and recording every skip — so the same job can be replayed without a model guessing on the happy path.",[190,1158,1159,1163],{},[199,1160,1162],{"href":394,"rel":1161},[203],"McKinsey’s State of AI"," keeps finding the same pattern: adoption is widespread, and operational redesign is not. Most organisations use a model somewhere. Few have redesigned how recurring work runs when nobody is watching. Loop engineering is that redesign for the boring path — the job that already has a checklist, a spreadsheet, and a named owner, but still lives in chat and memory.",[190,1165,1166],{},"Three nouns get conflated in every vendor conversation. Separate them before you compile anything:",[257,1168,1169,1177,1187],{},[260,1170,1171,1174,1175,306],{},[194,1172,1173],{},"Eval loop"," — a sensor that grades a run after the fact. Did the write match the signed payload? Did the test pass? See ",[199,1176,535],{"href":534},[260,1178,1179,1182,1183,306],{},[194,1180,1181],{},"Standing-order loop"," — a job that already exists before a run starts. A trigger admits the run; the known steps execute; a run record shows outputs and skips. See ",[199,1184,1186],{"href":1185},"what-is-a-nimbus-loop","what is a standing-order loop",[260,1188,1189,1192,1193,306],{},[194,1190,1191],{},"Agentic workflow"," — an interactive path that still reasons step by step when the route is not fully known. See ",[199,1194,1195],{"href":556},"what is an agentic workflow",[190,1197,1198,1199,1202,1203,1206,1207,1210,1211,306],{},"Loop engineering is none of those objects. It is the craft of making the ",[194,1200,1201],{},"known"," path explicit, cheap, and auditable. ",[199,1204,1205],{"href":238},"Harness engineering"," is the parallel craft for the ",[194,1208,1209],{},"model"," path: tools, hooks, and sensors around non-deterministic steps. You need both. They solve different failure modes. Compare ",[199,1212,1214],{"href":1213},"loop-engineering-vs-harness-engineering","loop engineering vs harness engineering",[252,1216,1218],{"id":1217},"what-is-loop-engineering","What is loop engineering?",[190,1220,1221,1222,1225,1226,1229],{},"Loop engineering treats recurring operational work as something you ",[194,1223,1224],{},"compile",", not something you ",[194,1227,1228],{},"re-describe"," every time.",[190,1231,1232,1235],{},[194,1233,1234],{},"Compile the known path."," Take the job that already has a finish line — nightly pipeline hygiene, weekly exception report, month-end accrual list — and turn it into a fixed sequence: read these sources, apply these rules, produce these outputs, notify these people. Those are the known steps. It should not depend on someone typing “please do the usual” into a chat box.",[190,1237,1238,1241,1242,1245,1246,1249],{},[194,1239,1240],{},"Detect change."," Reality moves. A new field appears in the CRM. A vendor changes an export format. A policy adds a threshold. Loop engineering includes sensors that notice divergence before the output lands in the wrong inbox. That is adjacent to an ",[199,1243,1244],{"href":534},"eval loop",", but the emphasis is earlier: catch drift in the ",[194,1247,1248],{},"inputs and rules",", not only grade the final artefact.",[190,1251,1252,1255,1256,363,1259,767,1262,1265],{},[194,1253,1254],{},"Record skips."," When the path cannot proceed — missing file, ambiguous row, policy exception — the loop must ",[194,1257,1258],{},"stop",[194,1260,1261],{},"say why",[194,1263,1264],{},"leave a record",". Skips are data. A skip count trending up is how you know the compiled path is stale. Silently “doing your best” overnight is not loop engineering. That is hope.",[190,1267,1268,1272,1273,1276,1277,1279,1280,1282,1283,306],{},[199,1269,1271],{"href":201,"rel":1270},[203],"LangChain"," defines Agent = Model + Harness. Loop engineering lives ",[194,1274,1275],{},"outside"," that equation on the happy path: no model invocation when the path is known. When the path is unknown, you escalate — to a person, or to an ",[199,1278,557],{"href":556}," inside a proper ",[199,1281,249],{"href":248},". See ",[199,1284,1286],{"href":1285},"a-loop-is-not-an-agent","a loop is not an agent",[1288,1289,1293],"pre",{"className":1290,"code":1291,"language":1292,"meta":171,"style":171},"language-mermaid shiki shiki-themes github-light github-dark","flowchart LR\n  compile[\"Write the known steps\"] --> wait[\"Wait for a trigger\"]\n  wait --> check{\"Does the world match?\"}\n  check -->|yes| run[\"Run the known steps\"]\n  check -->|no| skip[\"Skip and say why\"]\n  run --> record[\"Leave a run record\"]\n  skip --> record\n","mermaid",[320,1294,1295,1303,1308,1313,1319,1325,1331],{"__ignoreMap":171},[1296,1297,1300],"span",{"class":1298,"line":1299},"line",1,[1296,1301,1302],{},"flowchart LR\n",[1296,1304,1305],{"class":1298,"line":172},[1296,1306,1307],{},"  compile[\"Write the known steps\"] --> wait[\"Wait for a trigger\"]\n",[1296,1309,1310],{"class":1298,"line":888},[1296,1311,1312],{},"  wait --> check{\"Does the world match?\"}\n",[1296,1314,1316],{"class":1298,"line":1315},4,[1296,1317,1318],{},"  check -->|yes| run[\"Run the known steps\"]\n",[1296,1320,1322],{"class":1298,"line":1321},5,[1296,1323,1324],{},"  check -->|no| skip[\"Skip and say why\"]\n",[1296,1326,1328],{"class":1298,"line":1327},6,[1296,1329,1330],{},"  run --> record[\"Leave a run record\"]\n",[1296,1332,1334],{"class":1298,"line":1333},7,[1296,1335,1336],{},"  skip --> record\n",[190,1338,1339,1344,1345,1348],{},[199,1340,1343],{"href":1341,"rel":1342},"https:\u002F\u002Fwww.thoughtworks.com\u002Finsights\u002Farticles\u002Foperating-system-enterprise-ai",[203],"Thoughtworks’ operating system for enterprise AI"," splits harness layers from organisational ownership. Loop engineering is the operator layer for ",[194,1346,1347],{},"unattended"," work: who owns the known steps, who gets paged on skips, how exceptions re-enter the human queue. It is not layer 1 (pick a model). It is whether Tuesday’s job is the same object as Thursday’s.",[252,1350,255],{"id":254},[257,1352,1353,1359,1369,1375,1384,1390,1398],{},[260,1354,1355,1358],{},[194,1356,1357],{},"Known path."," The fixed steps for a job you already know how to do. Not a prompt. Not “figure it out.” A sequence you can replay.",[260,1360,1361,1364,1365,306],{},[194,1362,1363],{},"Trigger."," What admits a run: schedule, file arrival, threshold, or a deliberate manual start. See ",[199,1366,1368],{"href":1367},"six-things-that-start-a-loop","six things that start a loop",[260,1370,1371,1374],{},[194,1372,1373],{},"Skip."," A deliberate halt when the path cannot proceed safely. Recorded, not hidden.",[260,1376,1377,1380,1381,1383],{},[194,1378,1379],{},"Standing order."," A loop that waits for its trigger on a ",[199,1382,561],{"href":361},". The job exists before the run starts.",[260,1385,1386,1389],{},[194,1387,1388],{},"Run page."," One record per execution: inputs observed, steps taken, outputs attached, skips listed with reasons.",[260,1391,1392,1395,1396,306],{},[194,1393,1394],{},"Outputs vs actions."," Outputs are artefacts (report, list, draft). Actions are writes to live systems. Loop engineering defaults to outputs first; actions need ",[199,1397,524],{"href":366},[260,1399,1400,1403],{},[194,1401,1402],{},"Wizard, not chat ritual."," Operators start from a form with known fields, not an open-ended thread.",[190,1405,1406],{},"If a vendor uses “loop” for an eval sensor, a cron plus a prompt, or a multi-agent graph, ask which of the nouns above they actually mean. Vocabulary drift is how programmes buy the wrong control.",[252,1408,389],{"id":388},[190,1410,1411],{},"If recurring work still lives in direct messages and memory, you pay three taxes every quarter.",[190,1413,1414,1417],{},[194,1415,1416],{},"The reinterpretation tax."," Someone re-explains the job in Slack. Someone else runs a slightly different version. Audit asks what ran last month and the answer is a person’s name.",[190,1419,1420,1423],{},[194,1421,1422],{},"The overnight tax."," On-call gets paged. The only documentation is a clever prompt. The model guesses at ambiguous rows because nobody defined a skip.",[190,1425,1426,1429,1430,1434],{},[194,1427,1428],{},"The attribution tax."," Spend and outcomes cannot be tied to a named job. ",[199,1431,1433],{"href":394,"rel":1432},[203],"McKinsey"," shows usage without operational redesign; loop engineering is redesign for work that should not need a meeting to start.",[190,1436,400],{},[257,1438,1439,1442,1445,1456],{},[260,1440,1441],{},"the same report runs every week but only one person knows the steps",[260,1443,1444],{},"exceptions are handled in side threads with no record",[260,1446,1447,1448,1451,1452,1455],{},"auditors ask for evidence of ",[194,1449,1450],{},"what"," ran, not ",[194,1453,1454],{},"what was said"," in chat",[260,1457,1458,1459,1463],{},"you want ",[199,1460,1462],{"href":1461},"what-is-collaborative-ai","collaborative AI"," on outcomes, not personal copilots on fragments",[190,1465,1466,1471],{},[199,1467,1470],{"href":1468,"rel":1469},"https:\u002F\u002Faiindex.stanford.edu\u002F",[203],"Stanford HAI’s AI Index"," tracks capability and deployment. Capability rises faster than operational maturity. Loop engineering is maturity for the path you already know.",[190,1473,1474,1477],{},[199,1475,439],{"href":437,"rel":1476},[203]," Measure and Manage steps assume you can observe behaviour and change controls. A compiled loop with skip logs is observable. A chat thread is not.",[190,1479,1480,1483],{},[199,1481,445],{"href":443,"rel":1482},[203]," asks for documented operational controls and named actors. A standing order on a workstream with a roster beats a shared login to a consumer model.",[252,1485,1487],{"id":1486},"loop-engineering-vs-adjacent-crafts","Loop engineering vs adjacent crafts",[190,1489,1490,1493,1494,508,1496,306],{},[194,1491,1492],{},"Harness engineering."," Runtime around the model: tools, hooks, eval sensors. Necessary when the step requires judgement or generation. See ",[199,1495,1214],{"href":1213},[199,1497,1498],{"href":238},"what is harness engineering",[190,1500,1501,1504,1505,306],{},[194,1502,1503],{},"Eval loops."," Grade a run. Essential on write paths and agent steps. Not the same as admitting the run. See ",[199,1506,1508],{"href":1507},"how-to-evaluate-loop-engineering","how to evaluate loop engineering",[190,1510,1511,1514,1515,306],{},[194,1512,1513],{},"Workflow automation (RPA)."," Brittle screen replay. Loop engineering prefers API reads, explicit rules, and recorded skips over pretending the interface never changed. See ",[199,1516,1518],{"href":1517},"loop-vs-rpa","loop vs RPA",[190,1520,1521,1524,1525,306],{},[194,1522,1523],{},"Agentic workflow."," Reasoning path when the route is not fully known. Loop engineering handles the known segment; agentic workflow handles exploration. See ",[199,1526,1528],{"href":1527},"loop-vs-workflow-vs-agent","loop vs workflow vs agent",[190,1530,1531,1534,1535,1539],{},[194,1532,1533],{},"Chat ops."," Fast for novel problems. Expensive and un-auditable for recurring ones. ",[199,1536,1538],{"href":1537},"multiplayer-ai-and-multi-agent-ai","Multiplayer AI"," in a shared room is for collaboration; a loop is for repetition.",[190,1541,1542],{},"The test is simple. If a new hire can follow numbered steps and produce the same artefact, compile it. If the next step depends on reading the case, harness it. If you cannot tell which you are doing, you will agent-wrap a checklist and call it innovation.",[252,1544,1546],{"id":1545},"how-to-do-it","How to do it",[190,1548,1549,1552],{},[194,1550,1551],{},"1. Name one recurring job with a finish line."," Not “AI for finance.” “Weekly pipeline exception list for EMEA” is a job.",[190,1554,1555,1558],{},[194,1556,1557],{},"2. Write the path as steps a new hire could follow."," If the path cannot be written, it is not ready to compile. It is still tribal knowledge.",[190,1560,1561,1564,1565,306],{},[194,1562,1563],{},"3. Separate outputs from actions."," Default to read-only connectors and artefact outputs. Promote to signed writes only with ",[199,1566,1568],{"href":1567},"governance-as-a-multiplayer-primitive","governance as a multiplayer primitive",[190,1570,1571,1574],{},[194,1572,1573],{},"4. Define skips explicitly."," Missing source file → skip with reason. Ambiguous owner on row 47 → skip, assign to queue. Never “best effort fill.”",[190,1576,1577,1580,1581,508,1584,306],{},[194,1578,1579],{},"5. Attach the loop to a workstream."," Brief, connectors, roster, budget. See ",[199,1582,1583],{"href":361},"what is an AI workstream",[199,1585,1587],{"href":1586},"four-pillars-of-an-enterprise-ai-platform","four pillars of an enterprise AI platform",[190,1589,1590,1593],{},[194,1591,1592],{},"6. Review skip rates on a cadence."," Skips are product feedback. A rising skip count means the steps or the world changed.",[190,1595,1596,333,1599,1603],{},[194,1597,1598],{},"7. Keep agents off the happy path.",[199,1600,1602],{"href":1601},"agents-should-be-disposable","Agents should be disposable"," for exploration. The compiled loop should survive model churn.",[190,1605,1606],{},"A useful intermediate artefact is a one-page write-up of the known path: trigger, sources, rules, outputs, skip classes, named owner, named exception queue. If you cannot fill that page, you are not compiling — you are hoping a prompt will remember.",[252,1608,1610],{"id":1609},"a-worked-example-weekly-pipeline-exceptions","A worked example: weekly pipeline exceptions",[190,1612,1613],{},"Revenue operations runs the same exception pass every Monday: pull CRM opportunities in stale stages, join the account executive’s sheet, flag conflicts, send finance a list. Today it lives in one person’s notebook and a Slack reminder.",[190,1615,1616,1619],{},[194,1617,1618],{},"Compile."," Step one: read CRM with agreed filters. Step two: read the sheet from the shared drive path. Step three: apply conflict rules documented in the wiki. Step four: produce a table artefact. Step five: notify finance and the regional lead on the workstream roster.",[190,1621,1622,1624],{},[194,1623,1240],{}," If the sheet tab name changes, skip — do not guess. If CRM adds a required field, the sensor fires before rows silently drop.",[190,1626,1627,1629],{},[194,1628,1254],{}," Row with two owners → skip row, add to human queue. Missing sheet → skip run, page on-call. Every skip appears on the run page.",[190,1631,1632,1633,1636,1637,1639],{},"No model on the happy path. If a row needs judgement (“is this deal actually committed?”), the loop ",[194,1634,1635],{},"escalates"," to a person on the roster — or to an ",[199,1638,557],{"href":556}," if you later choose to add one for that class only. The default remains: do not guess.",[190,1641,1642,1643,1645,1646,508,1650,1654],{},"This is ",[199,1644,1462],{"href":1461}," for revenue operations without turning Monday into a chat ritual. Compare ",[199,1647,1649],{"href":1648},"loops-for-revenue-operations","loops for revenue operations",[199,1651,1653],{"href":1652},"collaborative-ai-for-finance-and-planning","collaborative AI for finance and planning"," when the artefact is a forecast pack.",[190,1656,1657,1661],{},[199,1658,1660],{"href":286,"rel":1659},[203],"Anthropic’s guidance on building effective agents"," warns against agentic complexity where a fixed pipeline suffices. Loop engineering is the operational form of that advice: use the simplest architecture that completes the job.",[190,1663,1664,1665,1670],{},"Research on human–AI teams (",[199,1666,1669],{"href":1667,"rel":1668},"https:\u002F\u002Fwww.nature.com\u002Farticles\u002Fs41562-022-01358-0",[203],"Yang et al., Nature Human Behaviour, 2022",") shows performance gains when roles are clear and humans retain override. Loops encode that split: machine for the known path, human for the exception queue.",[190,1672,1673],{},"Walk the same Monday after compile. The trigger fires at 06:00. The loop reads 1,842 opportunities and 31 sheet rows. It emits 14 exceptions and 3 skips (two dual-owned rows, one missing close date). Finance opens the run page, not a forwarded screenshot. The dual-owned rows sit in a queue with a reason. Nobody re-typed “please do the usual.” That is the job, compiled.",[252,1675,1677],{"id":1676},"what-goes-wrong","What goes wrong",[190,1679,1680,1683],{},[194,1681,1682],{},"Compiling a wish."," “AI for close” is not a known path. If you cannot name the sources and the finish line, you are still in discovery. Run the job manually twice and write what you actually did.",[190,1685,1686,1689],{},[194,1687,1688],{},"Best-effort fills."," A missing owner becomes “unassigned” or a guessed name. That is an agent decision without a harness. It violates the skip rule and teaches operators that silence means success.",[190,1691,1692,1695],{},[194,1693,1694],{},"Chat as the compiler."," A long system prompt is not a version of the known path. Prompts drift. New hires cannot start the job from a prompt they have never seen.",[190,1697,1698,1701,1702,306],{},[194,1699,1700],{},"Eval as admission."," A sensor that grades runs does not start them. You still need a standing order. See ",[199,1703,535],{"href":534},[190,1705,1706,1709,1710,306],{},[194,1707,1708],{},"Actions before outputs."," Writing to CRM in week one hides a bad path inside live records. Prove the table against the manual baseline first. Then quote a write. See ",[199,1711,1712],{"href":366},"what is write-back governance",[190,1714,1715,1718],{},[194,1716,1717],{},"Ownerless paths."," If nobody reviews skip trends, the compiled path rots. The world changes; the steps do not; skip rates climb; people go back to Slack.",[190,1720,1721,1724,1725,306],{},[194,1722,1723],{},"Agent-wrapping the checklist."," Every Monday a model re-reads instructions and re-derives spreadsheet logic. Cost scales with repetition. Drift scales with temperature. Audit sees a transcript, not which version of the steps ran. See ",[199,1726,1286],{"href":1285},[190,1728,1729,1732,1733,306],{},[194,1730,1731],{},"RPA theatre."," Replaying clicks looks like a compiled path until the interface changes. Prefer APIs and explicit skips. See ",[199,1734,1518],{"href":1517},[190,1736,1737],{},"Failure looks like a green check and a wrong list. Success looks like a skip with a reason and a human on the roster.",[252,1739,1741],{"id":1740},"governance-and-regulation-plain-english","Governance and regulation (plain English)",[190,1743,1744],{},"Loop engineering supports controls auditors already ask for, without requiring you to become a lawyer.",[190,1746,1747,1752],{},[199,1748,1751],{"href":1749,"rel":1750},"https:\u002F\u002Fwww.govinfo.gov\u002Fcontent\u002Fpkg\u002FPLAW-107publ204\u002Fhtml\u002FPLAW-107publ204.htm",[203],"Sarbanes-Oxley"," cares about trails for financial reporting. A compiled loop that produces a signed exception list — with skips recorded — is easier to walk than “the model said it looked fine.”",[190,1754,1755,1756,508,1761,1766],{},"The ",[199,1757,1760],{"href":1758,"rel":1759},"https:\u002F\u002Fartificialintelligenceact.eu\u002F",[203],"EU AI Act",[199,1762,1765],{"href":1763,"rel":1764},"https:\u002F\u002Fgdpr.eu\u002F",[203],"GDPR"," stress purpose limitation and documentation. A standing order on a scoped workstream states purpose (this job), data (these connectors), and actors (this roster). A god workspace does not.",[190,1768,1769,1773,1774,1777],{},[199,1770,1772],{"href":1771},"what-is-human-in-the-loop-ai","Human-in-the-loop"," remains mandatory for writes and for judgement calls. Loop engineering makes the ",[194,1775,1776],{},"automatic"," segment automatic; it does not remove named signers from actions.",[190,1779,1780],{},"Regulators rarely use the phrase “loop engineering.” They ask operational questions: who decided, what ran, can you replay it. A compiled path with a run page answers those questions for unattended segments. A chat log answers them poorly.",[252,1782,1784],{"id":1783},"how-to-start-this-quarter","How to start this quarter",[190,1786,1787,1788,1791],{},"Pick ",[194,1789,1790],{},"one"," recurring job. Not three. One.",[1793,1794,1795,1801,1809,1815,1821,1827,1833],"ol",{},[260,1796,1797,1800],{},[194,1798,1799],{},"Document the path"," as numbered steps. If step three is “ask the person who always knows,” stop — that step is a skip queue, not a step.",[260,1802,1803,1806,1807,306],{},[194,1804,1805],{},"Create a workstream"," with the brief, read-only connectors, and the roster who cares about the output. See ",[199,1808,1583],{"href":361},[260,1810,1811,1814],{},[194,1812,1813],{},"Implement outputs only"," for the first four weeks. No CRM writes. No journals.",[260,1816,1817,1820],{},[194,1818,1819],{},"Run on schedule"," twice alongside the legacy process. Diff the artefacts, not the vibes.",[260,1822,1823,1826],{},[194,1824,1825],{},"Add one skip rule"," you wish you had last month. Missing file, ambiguous row, stale tab name.",[260,1828,1829,1832],{},[194,1830,1831],{},"Review skip count"," in the fifth week. Update the steps or the wiki — that is loop engineering, not prompt tuning.",[260,1834,1835,1838,1839,1841],{},[194,1836,1837],{},"Decide whether any exception class"," deserves an ",[199,1840,557],{"href":556},". Most will not, yet.",[190,1843,1844,1847],{},[199,1845,1846],{"href":1507},"How to evaluate loop engineering"," is the buying sheet: can you compile, detect, skip, and replay without opening chat?",[190,1849,1850],{},"Score any product the way you score an operations tool: same job, same output, recorded skips, named humans on exceptions. If the demo starts in a chat box, you are scoring a copilot.",[252,1852,1854],{"id":1853},"how-this-shows-up-in-nimbus","How this shows up in Nimbus",[190,1856,1857,1858,1861,1862,1864,1865,1867],{},"The platform names a compiled standing order a ",[194,1859,1860],{},"Loop",". It lives on a ",[199,1863,561],{"href":361},": brief, connectors, roster, and ",[199,1866,382],{"href":366},". Operators configure the trigger and the known steps from a wizard; each execution gets a run page with outputs and skips.",[190,1869,1870,1871,1873,1874,306],{},"Loop engineering is the practice around that object — compiling the path, defining skip classes, reviewing skip rates — not a button labelled engineering. Writes still follow ",[199,1872,524],{"href":366},". Model steps, if you attach any, sit under ",[199,1875,1876],{"href":238},"harness engineering",[190,1878,1879,1880,363,1883,767,1885,1887,1888,508,1890,306],{},"See ",[199,1881,1882],{"href":1185},"what is a Nimbus Loop",[199,1884,1214],{"href":1213},[199,1886,1286],{"href":1285},". Product surfaces: ",[199,1889,362],{"href":32},[199,1891,382],{"href":40},[252,1893,758],{"id":757},[257,1895,1896,1901,1906,1911,1915],{},[260,1897,1898],{},[199,1899,1900],{"href":1185},"What is a Nimbus Loop",[260,1902,1903],{},[199,1904,1905],{"href":1213},"Loop engineering vs harness engineering",[260,1907,1908],{},[199,1909,1910],{"href":1285},"A loop is not an agent",[260,1912,1913],{},[199,1914,239],{"href":238},[260,1916,1917],{},[199,1918,1919],{"href":361},"What is an AI workstream",[252,1921,774],{"id":773},[257,1923,1924,1929,1935,1941,1947,1952,1958,1963,1968,1974,1979],{},[260,1925,1926],{},[199,1927,836],{"href":394,"rel":1928},[203],[260,1930,1931],{},[199,1932,1934],{"href":201,"rel":1933},[203],"LangChain, Agents",[260,1936,1937],{},[199,1938,1940],{"href":1341,"rel":1939},[203],"Thoughtworks, The operating system for enterprise AI",[260,1942,1943],{},[199,1944,1946],{"href":437,"rel":1945},[203],"NIST AI RMF",[260,1948,1949],{},[199,1950,445],{"href":443,"rel":1951},[203],[260,1953,1954],{},[199,1955,1957],{"href":1468,"rel":1956},[203],"Stanford HAI, AI Index",[260,1959,1960],{},[199,1961,812],{"href":286,"rel":1962},[203],[260,1964,1965],{},[199,1966,1669],{"href":1667,"rel":1967},[203],[260,1969,1970],{},[199,1971,1973],{"href":1749,"rel":1972},[203],"Sarbanes-Oxley Act of 2002",[260,1975,1976],{},[199,1977,1760],{"href":1758,"rel":1978},[203],[260,1980,1981],{},[199,1982,1765],{"href":1763,"rel":1983},[203],[1985,1986,1987],"style",{},"html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}html.dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}",{"title":171,"searchDepth":172,"depth":172,"links":1989},[1990,1991,1992,1993,1994,1995,1996,1997,1998,1999,2000,2001],{"id":1217,"depth":172,"text":1218},{"id":254,"depth":172,"text":255},{"id":388,"depth":172,"text":389},{"id":1486,"depth":172,"text":1487},{"id":1545,"depth":172,"text":1546},{"id":1609,"depth":172,"text":1610},{"id":1676,"depth":172,"text":1677},{"id":1740,"depth":172,"text":1741},{"id":1783,"depth":172,"text":1784},{"id":1853,"depth":172,"text":1854},{"id":757,"depth":172,"text":758},{"id":773,"depth":172,"text":774},"2026-09-12","loop engineering","Loop engineering is the discipline of compiling the known path for unattended work, detecting when reality diverges, and recording skips — not the runtime around a model, and not an eval sensor.",{"eyebrow":2006,"title":2007},"Common questions","Loops, harnesses, and sensors",[2009,2012,2015],{"question":2010,"answer":2011},"Is loop engineering the same as harness engineering?","No. Harness engineering designs the runtime around a model — which tools it may call, which hooks refuse a bad action, and which sensors grade the claim after the fact. Loop engineering compiles the known path for work that should not need a model on the happy path, and records when that path cannot proceed. Use harness engineering when a step requires generation or judgement. Use loop engineering when the steps are already known and the risk is silent drift rather than a clever prompt. Start with the compiled path if a checklist already exists. Refuse to treat hooks and eval sensors as a substitute for numbered steps you can replay without tokens.",{"question":2013,"answer":2014},"Is an eval loop the same as a Nimbus Loop?","No. An eval loop is a sensor that grades a run after the fact — tests, schemas, read-backs, signer checks. A standing-order loop admits a run when a trigger fires and then executes the known steps. Confusing the two is how programmes buy a grading harness and still start Monday's job from Slack. Use an eval loop on writes and on any model step that needs independent verification. Use a standing order when the job should exist before anyone types. Refuse to treat a score as proof that the same job ran, or a trigger as proof that the output was correct. See eval loops for enterprise agent harnesses and what is a Nimbus Loop.",{"question":2016,"answer":2017},"Do we still need agents if we have loop engineering?","Yes, for the unknown path. Loop engineering makes the known path cheap and unattended: read, rule, output, notify. When a row is ambiguous or a source is missing, the compiled path should skip — not invent. Agents and agentic workflows exist for exploration and judgement when the route is not fully known. Attach them only to an explicit exception class, inside a harness with stops and sensors. Refuse to put a model on the happy path of a job that already has a checklist, and refuse to leave exceptions as silent best-effort fills overnight.","\u002Fblog\u002Fwhat-is-loop-engineering",{"title":1145,"description":2004},"blog\u002Fwhat-is-loop-engineering",[901,2022,2023,2024,2025],"loop-engineering","loops","enterprise-ai","operations","qhxKV2_rt_cAEzSvf-oskijE-yUyn8jE5GsfZpSXdmM",{"enabled":165,"message":2028,"linkLabel":79,"linkHref":80,"id":2029,"title":2030,"archived":165,"authors":166,"badge":166,"body":2031,"date":166,"definedTerm":166,"department":166,"description":171,"extension":174,"eyebrow":166,"faqHeader":166,"faqs":166,"footerBand":166,"headline":166,"image":166,"industry":166,"jobType":166,"listed":131,"location":166,"navigation":131,"openRoles":166,"pageLayout":166,"path":2035,"relatedHeading":166,"seo":2036,"series":166,"sitemap":165,"status":166,"stem":2037,"subhead":166,"tags":166,"video":166,"whyJoin":166,"workplaceType":166,"__hash__":2038},"We're hiring! Join the team building the Sentient Enterprise.","content\u002Fshared\u002Fhiring.md","Hiring banner",{"type":168,"value":2032,"toc":2033},[],{"title":171,"searchDepth":172,"depth":172,"links":2034},[],"\u002Fshared\u002Fhiring",{"title":2030,"description":171},"shared\u002Fhiring","1zs3boivKda1e-b-hAyuNcmZSKjZUAXmecnwHVgcHzk",{"fold":2040,"id":2044,"title":2045,"archived":165,"authors":166,"badge":166,"body":2046,"date":166,"definedTerm":166,"department":166,"description":171,"extension":174,"eyebrow":166,"faqHeader":166,"faqs":166,"footerBand":2050,"headline":166,"image":166,"industry":166,"jobType":166,"listed":131,"location":166,"navigation":131,"openRoles":166,"pageLayout":166,"path":2054,"relatedHeading":166,"seo":2055,"series":166,"sitemap":165,"status":166,"stem":2056,"subhead":166,"tags":166,"video":166,"whyJoin":166,"workplaceType":166,"__hash__":2057},{"headline":2041,"description":2042,"primaryLabel":8,"primaryTo":2043,"secondaryLabel":924,"secondaryTo":12},"Run frontier AI your business actually owns.","Governed agent swarms, 2,000+ integrations, and a knowledge graph that stays inside your walls. Start on Free.","\u002Fsignup?plan=free","content\u002Fshared\u002Fcta.md","Site CTAs",{"type":168,"value":2047,"toc":2048},[],{"title":171,"searchDepth":172,"depth":172,"links":2049},[],{"headline":2051,"description":2052,"primaryLabel":8,"primaryTo":2043,"secondaryLabel":2053,"secondaryTo":85},"See what governed AI looks like on your stack.","Connect your tools, run a workstream, and keep every decision on your ledger. Start on Free.","Talk to our team","\u002Fshared\u002Fcta",{"title":2045,"description":171},"shared\u002Fcta","PS2VPJsszmUpMBZT6nEp8cWXCdeiN6zDRl-p8d0uY2k",1790215700458]