A vendor demos something shiny. A competitor posts about their "AI transformation." So you greenlight a project, and six months later you have a chatbot nobody uses and a bill nobody enjoyed. The technology was never the problem. The direction was.
The simple version
AI consulting is not building the AI. It is the part that happens before building, where we work out which problems in your business are actually worth pointing AI at, and which are cheaper to leave alone. You end with a short, ranked list and a costed plan. That is the whole job.
The analogy: a map before the trip
You would not start a long drive by flooring it in a random direction and hoping. You open a map, pick the destination, then choose the route that avoids the traffic. AI strategy is the map. The build is the drive. People who skip the map drive fast and arrive nowhere.
How a strategy engagement actually runs
Four steps, no mystery. Each one exists to stop you spending big before you know the bet is real.
Where AI usually pays off (and where it doesn't)
The honest map has a "skip this" branch. That is the part most sales decks hide. A good strategy says no to most ideas so the few that survive actually get built well.
Why it matters
Most AI budgets are not lost to bad technology. They are lost to good technology pointed at the wrong problem. An afternoon spent drawing the map saves the quarter spent driving the wrong way. We diagnose first, then build. That order is the whole point.