The Technology Works. The Organisation Doesn't Change Around It.

Most Artificial Intelligence projects don't fail because the technology doesn't work. They fail because the organisation doesn't change around it.

I've been studying this in detail lately, exploring how innovation theory applies specifically to AI, and one idea from the academic literature keeps hitting closest to the reality I see with clients “Organisations are bundles of routines.” Approval workflows. Decision rights. How work moves between people. What gets measured. These routines are incredibly stable and they're designed to be. So, when you drop an AI model into an organisation without touching them, very little actually changes. The AI saves the analyst two hours but those two hours flow nowhere and the transformation on the slide deck stays on the slide deck.

This is why the question I've started asking leadership teams isn't "which AI tools are you deploying?" It's: "What routines have changed because of AI?"

If the answer is none (or vague), the investment isn't yet creating value.

There are many places where organisational routines need to change when AI arrives, and a full operating model transformation touches far more than four. But in my experience, these are the most immediate the ones that tell you fastest whether an AI deployment is actually creating value or just creating activity:

• Decision rights: who approves AI output, and when does it go straight to action? 

• Workflows: which steps in the process does the AI now own end-to-end? 

• Measures: are you measuring outputs that reflect AI's actual contribution, or just activity?

• Skills: does your team know how to work alongside AI, or just how to use it?

This is week one of the next six weeks studying innovation and then applying it through an AI lens. What I'm finding is that the frameworks aren't new, but the urgency is.

What's your experience? Have you seen AI investments land without the organisational change to match?