Every business is being told to use AI, as though AI were a place to arrive at. It is not a strategy. It is a lever, and a lever does nothing on its own. It only matters once it is attached to something worth moving.
Point AI at a clear problem with a measurable outcome and it is extraordinary. Point it at nothing in particular and it produces impressive, confident output that changes no number in the business. The aim has to come first, and the aim is the hard part. The leverage was never the bottleneck.
AI laid over a good process makes it faster and cheaper. AI laid over a broken process makes the brokenness faster and cheaper too. It does not repair the system, it scales whatever is already there. If the thing underneath is wrong, more leverage is the last thing it needs.
Where is the business actually leaking. What would closing that gap be worth. What is the simplest thing that closes it. Sometimes the answer leans heavily on AI. Sometimes it does not touch it. The discipline is identical either way, because the discipline was never about the tool.
Because AI has made building cheap, it is tempting to build for the sake of it. The constraint has moved. It is no longer cost, it is judgement: knowing what is worth building at all. That is the scarce thing now, and no model supplies it for you.
I treat AI as one of the sharpest tools available, and still start exactly where I always have. With the gap, and the number.
That is the question before any build.