Most AI Strategies Start With the Wrong Question

Most AI strategies don't fail at execution. They fail at the first question.

When pressure builds — from the board, from competitors, from the headlines — leadership teams move fast to "do AI."

So they ask: where can we apply this?

It sounds reasonable. It's the wrong question.

It pulls the conversation toward tools before the problem is defined. And that's where strategies quietly unravel.

The organisations that get real value from AI start somewhere different.

They look at the last three years of board packs and QBRs. They find the problems that appear every year — the ones that never seem to get fixed. Revenue leakage. Slow decisions. Operational drag. Customer friction that never quite gets resolved.

Then they ask: if AI could help us solve one of these, what would that actually be worth?

That reframe changes the whole conversation. Instead of evaluating tools, you're designing solutions. Instead of "where does this fit?", the question becomes "how much is this worth if it works?"

We've seen this shift take leadership teams from scattered experimentation to a funded programme in a matter of weeks.

The technology is rarely the constraint.

The question usually is.