The strategy was right. Nobody could build it.

The deck is excellent. You've read it, or one like it: the durable advantage in AI isn't the model — those are commoditizing by the quarter — it's proprietary intelligence. Your own data, your own workflows rebuilt from the core, a learning loop where evals and monitoring are the real work and never stop. Owned in-house. Undelegable. It sits with the CEO. Every line of it is true, and a top-three strategy firm just put it in front of every board you know.

Then the engagement ends, the team rotates off, and you're holding a beautifully argued thesis about a system that still doesn't exist.

The pattern underneath

Naming the moat and building it are two different jobs, and they reward opposite things. A strategy firm's product is the recommendation — the framework, the board-ready deck — and it is priced to grow and built to end. "Rebuild your workflows from the core" becomes a multi-quarter program. "Evals are sixty percent of the work and monitoring never stops" becomes a slide, and the slide ships on a date. That is not a failure of the firm; it is what the product is for. The gap is that the moat it describes is built by staying, and a recommendation is finished when it is delivered. The advice is sound. Somebody still has to be there in month nine.

The reframe is simple once you see it: the thing of value is the working intelligence, not the document describing it. And a working system needs the senior hand on the build, not just in the room for the readout — because the judgment that diagnoses the leak is the same judgment that has to be encoded into the agent, tuned through the evals, and wired into the monitoring that keeps it honest. Hand that off to a junior and you get the artifact a junior can produce: a deck. Keep it on the senior, and make AI the leverage instead of an army, and what lands is the moat itself — built, owned, and still running after the invoice clears.

So keep the deck — it's a good map. Just don't confuse it with the moat. The framework is right that the advantage is built, owned, and undelegable. The part it can't say from inside a strategy model is who builds it. That seat — the senior who diagnoses the leak and then ships the fix the team keeps — is the whole job. It's the half the deck can't do.