AI as a stress test

For years before Covid, organisations debated whether working from home could work. There were pilots, strong opinions and plenty of reasons why it might not scale. Then Covid arrived, and the debate effectively ended. Not because the question had been resolved in theory, but because circumstances forced a system-wide stress test.

What that period revealed was interesting; organisations that already had clear decision-making, trust in their teams, and work designed around outcomes adapted far more easily than those relying on proximity, informal escalation, or managerial oversight to keep things moving. Covid didn’t create those differences, rather it exposed them.

I think AI is playing a similar role now. For a long time, many organisations have known that they aren’t especially well optimised for flow. Decisions take longer than they should, ownership is sometimes ambiguous, and dependencies are often tolerated rather than deliberately designed. Progress still happens, but it is frequently held together by experience, goodwill, and quiet heroics.

As AI introduces faster feedback loops, automation, and agentic behaviour into these environments, that underlying reality becomes harder to ignore. The slack that once absorbed friction disappears, and long-standing issues surface quickly. In that sense, AI isn’t the transformation, it is the stress test that reveals where organisational change has been overdue for some time.

Over the next few posts, I want to explore AI less as a destination and more as a diagnostic, looking at what this new pressure reveals about decision-making, dependencies, ownership, and how teams are actually designed to work. For me its less about how AI will change our organisations and more about what it’s already telling us about the ones we have.