The leapfrog test

Boards keep asking the same question — “are we doing enough on AI?” — and getting the same useless answer: a list of pilots. Pilots tell you what you’re doing. They don’t tell you whether it matters. For that you need to look from the outside in.

Here’s an exercise we run early in most engagements. It takes an afternoon, needs no technology, and is usually the moment the room goes quiet.

The exercise

Imagine a well-funded, AI-native competitor launching into your market next quarter. No legacy systems, no headcount to protect, no process to defend. They’ve read the same research you have, but they’re building from zero around it.

Now answer three questions about your single most profitable workflow:

1. What would they not build? Which steps, teams, or handoffs simply wouldn’t exist in their version?
2. What would they charge? If their cost to deliver is a fraction of yours, what does that do to your pricing — and your margin?
3. How long until your customers notice? Not “could they switch,” but “when the difference becomes obvious, how fast does it move?”

Reading the result

If the honest answers are “not much,” “about the same,” and “years” — good. You have a genuine moat, and your AI work should be about widening it.

But for most companies, in their most profitable workflow, the answers are uncomfortable: a competitor would skip half the steps, charge meaningfully less, and the gap would become visible inside a year. That’s not a reason to panic. It’s a map.

The point of the leapfrog test isn’t fear. It’s to find the one place you’d be hit first — so you can build it before someone else does.

The companies that win the next few years aren’t the ones with the most pilots. They’re the ones who ran this test honestly, picked the workflow that came back worst, and rebuilt it themselves while they still had the lead.

That workflow — the one that just made the room go quiet — is where we’d start.