"AI consultant" might be the most overused job title of the decade. Behind the noise, though, sits a real and useful discipline — and if you run a business, it's worth knowing what it should look like.
Here's the plain-language version: what the work actually involves, what it isn't, and the questions that quickly separate the genuine article from the bandwagon.
Strip away the jargon and it's three jobs
Finding what's costing you. Before any technology enters the conversation, the real work is understanding how your business actually runs — not how the org chart says it runs. Where does work pile up? What gets re-typed three times? Which numbers does nobody quite trust? A good consultant spends their first efforts here, because everything else depends on it.
Redesigning how work flows. Most operational pain isn't a software problem — it's a process that grew by accident rather than design. The second job is straightening that out: removing duplication, getting knowledge out of people's heads, and making the way work moves through the business deliberate.
Building the fix. Only then does AI earn its place — automating the repetitive steps, connecting the systems that don't talk, putting live numbers in front of the people who decide. The technology is the means. The product is a business that runs better.
What it isn't
It isn't reselling software licences with a margin. It isn't bolting a chatbot onto your website and calling it transformation. And it isn't a 200-page strategy document that nobody opens twice. If an engagement ends with a deck instead of a working improvement, something has gone wrong.
What good looks like
You should expect plain language — if you can't explain the recommendation to your team, it isn't finished. You should expect numbers attached to everything: what the problem costs, what the fix returns, and how long it pays back. You should expect working software, not promises. And you should expect to own what's built — fully, with no licensing strings or quiet lock-in.
Five questions to ask anyone you're considering
1. "What will you do before you recommend anything?" If the answer isn't some version of understand your business first, keep looking.
2. "What will it cost — and when do I find out?" A fixed price, agreed in writing before work begins, protects you. An open-ended day rate doesn't.
3. "Who owns what you build?" The only good answer is: you do. Entirely.
4. "What happens when you leave?" Good work survives its maker — documented, handed over, and usable without them on speed dial.
5. "What happens if the numbers don't justify the work?" The honest answer — then we'll tell you not to do it — is the one you want to hear.
Ask those five and you'll learn more in ten minutes than any brochure will tell you. They're also, not coincidentally, the standards we hold our own work to.