{"slug": "ai-operating-model-why-scaling-ai-is-an-organisational-design-problem", "title": "AI Operating Model: Why Scaling AI Is an Organisational Design Problem", "summary": "A new analysis argues that scaling AI is fundamentally an organizational design problem, not just a technological one. The piece highlights a Deloitte report from 2026 showing that while many technology leaders believe they can deploy and govern AI at scale, nearly three-quarters expect their operating model to change within 12 to 18 months. It emphasizes the need for an AI operating model that connects technology, governance, and human judgment to turn experiments into durable capability.", "body_md": "AI adoption is getting easier. Scaling AI is not.\n\nModels are more capable. APIs are easier to access. Copilots can be deployed quickly. Teams can prototype useful workflows in days.\n\nYet many organisations still struggle to turn that activity into durable capability.\n\nThe reason is increasingly clear: **AI does not scale through technology alone. It scales through an operating model.**\n\nThat means deciding who owns AI, how use cases are selected, how risk is governed, how learning is shared, how human judgement stays in the loop, and how successful experiments become part of normal work.\n\nThis is why the phrase **AI operating model** is becoming more important. In 2026, Deloitte reported a striking gap: while many technology leaders believe they can deploy and govern AI at scale, nearly three-quarters still expect their operating model to change within 12 to 18 months. The problem is moving from “Can we use AI?” to “Can the organisation absorb it well?”\n\nThat is a different question.\n\nAn AI operating model is the organisational system that determines how AI decisions are made, governed, funded, built, adopted and improved.\n\nIt is not simply an AI strategy document. It is not an AI governance policy. And it is not a central AI team with a new name.\n\nA useful AI operating model connects at least six things:\n\nThe technology matters. But the model around the technology determines whether it becomes capability or remains experimentation.\n\nMost organisations do not have an idea shortage.\n\nThey have use-case lists.\n\nCustomer support wants summarisation. Finance wants forecasting. Marketing wants content acceleration. Product wants research synthesis. Engineering wants coding assistance. Operations wants automation. Leadership wants better decision support.\n\nThe common response is to launch pilots.\n\nPilots are useful because they lower the cost of learning. But when each team pilots independently, the organisation can create a new kind of fragmentation:\n\nThe result can look like “lots of AI” without much institutional capability.\n\nThis is where an AI operating model becomes useful. It creates a way for learning to compound rather than reset inside every team.\n\nOne way to read AI adoption is through three interacting lenses: **psychology, technology and organisations**.\n\nThat matters because AI changes all three at the same time.\n\nAn AI system can be technically excellent and still fail if people do not trust it, understand it or know when to override it.\n\nAdoption is shaped by questions such as:\n\nThis is why “training” is too narrow a word for AI adoption.\n\nThe real issue is behaviour.\n\nAn AI operating model has to create confidence without creating complacency. It should make good judgement easier, not merely make AI available.\n\nThe second layer is technical.\n\nAI needs access to data, tools, workflows and applications. As capability increases, so does the importance of architecture and controls.\n\nOrganisations therefore need clear answers to questions such as:\n\nThis is where standards and risk frameworks matter. NIST’s AI Risk Management Framework and its Generative AI Profile emphasise lifecycle risk management, while ISO/IEC 42001 treats AI as a management-system question rather than a one-off technical control.\n\nThose frameworks are useful because they reinforce a simple point: responsible AI requires repeatable organisational processes around the technology.\n\nThe third layer is organisational.\n\nSomeone has to own the decisions.\n\nThat sounds obvious, but AI often crosses existing boundaries. A customer-facing AI feature may involve product, technology, legal, security, data, operations and customer support. A coding assistant may affect engineering quality, intellectual property, security and productivity measurement. An internal agent may touch systems owned by several teams.\n\nTraditional organisational charts do not automatically resolve these questions.\n\nThe AI operating model therefore has to define:\n\nWithout this clarity, governance becomes a queue and delivery becomes negotiation.\n\nA common AI operating model question is whether AI should be centralised.\n\nThere are good reasons to centralise parts of it. Shared standards, architecture, security, evaluation, vendor decisions and governance can become expensive and inconsistent when every business unit rebuilds them independently.\n\nBut centralising every use case creates another problem: the people closest to the work lose ownership.\n\nThat is why many enterprise AI models are moving toward a **federated** structure.\n\nThe centre holds the things that should be common:\n\nBusiness and product teams hold the things that require context:\n\nThe centre should not become the place where every AI decision waits.\n\nIts job is to make good distributed decisions possible.\n\nThis is where a Centre of Excellence can help — if it is designed correctly.\n\nA weak AI CoE becomes a committee that reviews requests.\n\nA stronger one becomes an organisational mechanism for making AI capability repeatable.\n\nIts role may include:\n\nThe test is simple: **does the CoE make the organisation more capable without making it more dependent?**\n\nCralgo explores this broader capability question in its work on [Centres of Excellence](https://cralgo.com/framework) and [technology as an organisational system](https://cralgo.com/technology).\n\nAI governance is often discussed as a control problem.\n\nIt is also an execution-design problem.\n\nIf governance only happens at the end of a project, teams will either wait too long or work around it. If every use case receives the same review, low-risk experimentation becomes unnecessarily slow while genuinely important risks can receive too little attention.\n\nA better AI operating model makes governance proportional.\n\nFor example:\n\nA meeting-summary tool using approved enterprise data may require lightweight controls, clear retention rules and basic quality checks.\n\nAn AI system that recommends operational actions may require stronger logging, human approval and explicit rollback paths.\n\nAI used in areas such as healthcare, employment, financial decisions or safety-sensitive operations may require independent validation, documented evidence, tighter monitoring and formal accountability.\n\nThe point is not to make governance smaller.\n\nIt is to make governance fit the consequence of the decision.\n\nAs AI systems become more capable, organisations can be tempted to move more decisions into the system.\n\nBut capability and authority are not the same thing.\n\nA model may be able to recommend an action without being the right place to own that action.\n\nThat distinction becomes especially important with agents that can call tools, update systems, trigger workflows or communicate with customers.\n\nThe key design question changes from:\n\n**What can the AI do?**\n\nTo:\n\n**What should the AI be allowed to do, under what conditions, with whose judgement around it?**\n\nThis is one reason the human side of AI cannot be separated from the technical side. Trust, attention, decision-making and accountability all shape the outcome.\n\nBefore scaling AI across an organisation, leadership teams should be able to answer these questions clearly:\n\nIf these answers are vague, the organisation probably does not yet have an AI operating model. It has AI activity.\n\nThose are not the same thing.\n\nThe next phase of enterprise AI will not be decided only by who has access to the strongest model.\n\nAccess is becoming easier.\n\nThe harder advantage is organisational: the ability to decide well, deploy safely, learn quickly, distribute capability and retain judgement as AI becomes embedded in normal work.\n\nThat is why AI operating models matter.\n\nThe technology changes what becomes possible.\n\nPeople determine how it is understood and used.\n\nOrganisations determine whether that possibility becomes repeatable capability.\n\nThe outcome emerges from all three.\n\n**Cralgo** is a research and technology company exploring how psychology, technology and organisations shape better outcomes.\n\nExplore [Cralgo](https://cralgo.com), [Technology](https://cralgo.com/technology), and the [Operating Model](https://cralgo.com/work/operating-model).", "url": "https://wpnews.pro/news/ai-operating-model-why-scaling-ai-is-an-organisational-design-problem", "canonical_source": "https://dev.to/cralgo/ai-operating-model-why-scaling-ai-is-an-organisational-design-problem-2cg9", "published_at": "2026-09-01 05:01:42+00:00", "updated_at": "2026-09-01 05:21:40.872182+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-policy", "ai-ethics", "ai-infrastructure"], "entities": ["Deloitte", "NIST", "ISO/IEC 42001"], "alternates": {"html": "https://wpnews.pro/news/ai-operating-model-why-scaling-ai-is-an-organisational-design-problem", "markdown": "https://wpnews.pro/news/ai-operating-model-why-scaling-ai-is-an-organisational-design-problem.md", "text": "https://wpnews.pro/news/ai-operating-model-why-scaling-ai-is-an-organisational-design-problem.txt", "jsonld": "https://wpnews.pro/news/ai-operating-model-why-scaling-ai-is-an-organisational-design-problem.jsonld"}}