The first AI use case you find is the wrong place to start.
When something feels painful, the instinct is to fix it quickly. That instinct is what turns an AI programme into activity instead of leverage.
The instinct is understandable
In many of the conversations I have with leaders about AI, the focus goes straight to buying a tool, or to what I would call unstructured experimentation. Someone mentions an inefficiency in daily operations, a reporting bottleneck in finance, or content generation in marketing, and the instinct is to jump right in.
It is understandable. When something feels painful, the natural reaction is to try to fix it quickly.
Start with outcomes, not applications
Before we talk about tools or workflows, we need clarity on what the business is actually trying to achieve.
What has to be true twelve months from now for the board to say this was a good year? Which two or three metrics genuinely drive performance? Is the priority margin expansion, revenue growth, retention, productivity, or something else entirely?
Only once that is clear does it make sense to ask which behaviours must change and what capabilities are missing today. And only after that do use cases and prioritisation become meaningful.
Activity, or leverage
AI should strengthen the operating model you are building rather than run alongside it. In practice, the sequence matters more than the shortlist: outcomes first, then the behaviours that have to change, then the capabilities that are missing, and only then the use cases that serve them.
It is a slower start and a much shorter road.
Without strategic clarity, use cases are activity. With clarity, they become leverage.
If you are at the beginning of this, these are the pages that carry it into practice.