The short answer

How do I choose an AI consultant?

Ask three questions. What will we own at the end? Can we see something already running rather than a presentation? And what would you tell us not to build? The third one is the tell. A consultant who cannot name something they would talk you out of is selling hours rather than outcomes.

The questions that separate real from rehearsed

Ownership first, and get specifics rather than reassurance. Where does the code live, whose name is on the hosting account, can you export the data without asking, and what happens if you stop working together. Vague answers here are the single most reliable warning sign.

Then evidence. Anyone can show a deck of what is possible. Ask to watch something actually running, ideally something unglamorous, because a working invoice chaser is more informative than an impressive demonstration built for the meeting.

Then honesty about scope. Every business has things worth building and things a cheap tool already does well. A consultant who recommends building all of it has an incentive problem, and you will pay for it.

Warning signs

A proposal that starts with a platform rather than a problem. Pricing that scales with your headcount rather than with the work. Reluctance to name what the first project will measure. And any arrangement where stopping the relationship means losing the systems.

Also be careful with impressive breadth. A studio that claims equal depth in everything is describing a sales page rather than a team. It is more useful to hear plainly what they are good at and what they would bring somebody else in for.

What a good engagement feels like

The first conversation is mostly about your business rather than their technology, and you leave it with a clearer picture of your own operation whether or not you hire anyone.

The first project is narrow, measured, and live in weeks. And at every point you can see what has been built, where it runs, and what would happen if you walked away.

And then

What people ask next.

Should we hire a consultant or an agency?

The label matters less than what you end up owning and who does the work. Ask whether the people in the room are the people building it, and ask what the arrangement looks like in year two. A team that hands over working systems is a better fit than one that keeps you on a retainer to access your own tools.

Is it better to hire someone in house?

Eventually it often is, and a good outsider should be building toward that rather than against it. Hiring in house first is expensive and risky when nobody yet knows what the role should do. A narrow project usually answers that question cheaply, and then the hire is an informed one.

The answer to what you would not build is the one that matters.

Book an AI audit Every question

Built in Grand Rapids, Michigan, and put to work wherever your business is.