I love an axiom. Something easy to remember, fast to say, and punchy enough to stick. 

With my older boys, I’ve often said, “If in doubt, don’t.” With my younger son, who is autistic, I say, “Stay close, stay safe.” Short phrases. Big truths. The kind that helps in the moment when time is short and the stakes are high. 

That is probably why ideas like Moore’s Law, Amara’s Law and Parkinson’s Law continue to resonate with technology leaders. They help us hold on to simple ideas as we make sense of tech adoption, value, and implementation at scale. They are part of how we stay the course and keep our people anchored when the media shrill rises, the consultants’ decks get thicker, and LinkedIn pundits fill our feeds with certainty. 

And that brings me to Conway’s Law. 

Platform Choice Is Not The Starting Point 

I have been presenting multiple times per week to public sector clients a workshop session entitled “Overcoming The Hurdles Of AI Adoption At Scale”. Recently, a client said, “Conway’s law plays out every single time. We want to implement systems before addressing our business … and every time we end up with the same results, in that the systems end up just as messed up as our organisations.” In other words: start with the operating model and organizational structure, then orient platforms to the right domains. 

One of the great mistakes in AI right now is the belief that the answer lies primarily in choosing the right platform, model, or vendor stack. It does not. If the operating model is unclear, fragmented, or built for an earlier era of work, the AI system will inherit those flaws and reproduce them at machine speed. That is why Conway’s Law feels so relevant again: systems do not transcend organizations. They mirror them. And in the age of agentic AI, they amplify the worst of them. The silos, the politics, and more. 

Start With Your Organization And Your People 

That point sits at the heart of what we are doing with our research into the Cognitive Operating Model, Intelligence Enterprise, and Skills-Oriented Architecture. And the core premise of this research is the AI productivity paradox: gains dissipate inside operating models designed for human-only, task-based work. Bolting agents onto yesterday’s roles, workflows, and decision rights is technology deployment with better marketing from companies who need to maximize IPO valuation to get the capital needed to feed the AI cash furnace. 

That is also why the shift from generative AI to agentic AI matters so much. Generative AI was the warm-up. Agentic AI changes the game because we move from prompts to plans. These systems now retrieve, decide, trigger, notify, and act. That shifts the conversation from output quality to governance, accountability, orchestration, and legitimacy. And in government, directly into explainability, fairness, and public trust. 

The Operating Model Shift Matters 

If your operating model is siloed, fragmented, overloaded with handoffs, and built around a human-only conception of work, your AI estate will reflect that complexity. Agents will be selected, deployed, and governed according to those same fault lines. The result? duplicated capabilities, fragmented context, inconsistent controls, and point solutions masquerading as transformation. 

What Conway’s Law explains is why the operating model shift is so central. At its core, Agentic AI is a work architecture problem and it’s an operating model shock. If agents increasingly become the default executors of routine cognitive work, then the organization must be redesigned around that reality. Roles, workflows, escalation paths, management assumptions, and accountability models all change. Otherwise, the technology will simply automate the archaeology of today’s enterprises. 

The Skills And Context Matters 

This is why our work encourage our clients to move away from use-case thinking and toward skills as the atomic unit of design. 

A use case describes a problem to solve. A skill describes a bounded cognitive capability that can be reused, governed, and composed across roles and workflows. Organize agentic portfolios around isolated use cases and you get fragile, siloed deployments that resist scale. Organize around skills and you create the conditions for composition, governance, and durable operating-model change. Dynamic, agile, and flexible. 

The other half of this is context. Capability on its own is not enough. Real competence depends on the surrounding semantic layer of policy, vocabulary, memory, decision traces, tacit knowledge, and organizational logic.  Without a coherent way to surface and govern context, agentic systems will mirror the enterprise’s missing knowledge, fragmented policy interpretation, weak accountability, and rising costs. 

Conway’s Law Matters 

If I had to turn Conway’s Law into a practical checklist for leaders in the age of agentic AI, it would be this: 

  • Start with the operating model. Let the platform follow the work, the problem domains, and the outcomes the organization needs to achieve. 
  • Build reusable organizational capabilities. Design skills, roles, workflows, and governance structures that compound across use cases. 
  • Treat context as organizational intelligence. Make policy, knowledge, memory, and decision logic machine-readable, governable, and available at the point of work. 
  • Design agents around the organization you want to become. Agents amplify the system they operate within, including its strengths, gaps, and accountability model. 

For me, that is the modern value of Conway’s Law. In the dizzying storm of change we’re in, if we want agentic AI to create compound value, we must first redesign the operating model that surrounds it. That is the work. That is the hurdle. That is why our current research is so focused on structure, context, and the redesign of work itself. 

Otherwise, we are not building the future of work. We are automating the past. 

So, remember kids: “Operating Models Deliver Outcomes.”