The first question I ask every client — and it's never about AI

When someone comes to us for help with their AI strategy, the first question I tend to ask is a simple one: what problem are you trying to solve?

The answer tells me almost everything I need to know. If the response is clear and specific — we are losing time to a particular process, our customer data is too fragmented to act on, our team is spending hours on work that should take minutes — then we are in a good position. There is a defined problem, and the question is whether AI is the right solution to it.

If the answer is more like ‘we need to be doing more with AI’ or ‘the board has asked us to develop an AI strategy,’ then we are in a different conversation. Not a bad one, necessarily, but a more foundational one. Because the starting point for any meaningful AI deployment is not the technology — it is an honest understanding of what the business needs.

This matters for a practical reason. AI is genuinely useful across a remarkable range of contexts. But it is not useful in all of them. There are tasks where AI introduces more risk than it removes. There are problems where the real issue is process design, not technology. There are situations where the most valuable thing an advisor can do is tell a client that the AI project they have scoped is solving the wrong problem.

Sometimes a business has challenges around people that AI can’t fix or mask. Not enough people with a specific skill, or not the right one needed at this stage of the business’ life. Insufficient training, or managers who have been promoted because of their technical skills and not their ability to manage and lead — let’s never forget that management is not the same as leadership.

We have had conversations that started with a brief to build an AI deployment strategy and ended with a recommendation to fix the data architecture first. We have worked with organisations that came to us expecting to transform their operations with AI and left with a clearer sense of which two or three use cases would actually move the needle — and why the other fifteen on their list were distractions.

None of this requires cynicism about AI. Quite the opposite. Taking the technology seriously means being honest about where it works and where it does not. It means asking clients the questions that might complicate the brief, rather than the ones that make the engagement easier to scope. It means being willing to say, occasionally, that the most valuable thing we can do is stop you from spending money in the wrong place.

The best AI strategies we have seen are not the most ambitious ones. They are the ones that start with a clear problem, match the right solution to it, and build from there. The technology is capable of remarkable things — but only if you give it the right job to do.