What good looks like
A strong consultant starts with your problem, not their favorite tool. They ask what a win would actually change, how you'd measure it, and what could go wrong, before anyone talks models or budgets.
They're specific about scope, honest about what AI can and can't do reliably, and clear that you'll own the work when it's done. You should come away understanding the plan, not just trusting it.
- Leads with your business outcome, not a specific product
- Can show real work or walk you through a concrete approach
- Defines success metrics before starting
- Is candid about limits, failure modes, and what AI won't fix
- Leaves you with code, docs, and access you fully own
- Prices the work in a way you can understand and check
Warning signs to avoid
Be wary of anyone who guarantees a specific accuracy or revenue number before seeing your data, name-drops buzzwords instead of explaining the approach, or wants to lock you into a black box only they can maintain.
Also watch for vague scope, no plan for measuring results, and hand-waving about privacy and data handling. If they can't explain where your data goes and who can see it, that's a problem.
- Guaranteed results or 'best in the world' claims before any discovery
- Buzzwords with no plain-English explanation
- No clear success metric or way to test the outcome
- Lock-in: you can't run or move the solution without them
- Vague answers on where your data lives and who can access it
- Scope that keeps expanding with no fixed deliverable
Questions worth asking on the call
A few direct questions separate real practitioners from resellers. Ask them to describe a project that didn't go to plan and what they changed, and ask what they'd need to see in your data before promising anything.
The best answers are specific and occasionally admit uncertainty. That honesty is a feature, not a weakness.
- What would make you tell us not to build this?
- How will we measure whether it worked?
- What happens to our data, and where does it run?
- What do we own and can maintain when you're done?
- Walk me through a project that went sideways
More on ai strategy & consulting
Frequently asked questions
Should I hire a consultant or just try tools myself first?
Try the easy tools yourself first. Bring in outside help when you hit private data, system integrations, high volume, or a decision that's expensive to get wrong. A good consultant will actually tell you when you don't need one yet.
How much AI experience should a consultant have?
Depth matters less than judgment. Look for someone who has shipped real, working software, understands your kind of business, and can explain trade-offs honestly. Beware anyone whose only proof is a slide deck full of logos.
What does a first engagement usually look like?
Often a short discovery or workflow review to find where AI actually helps, followed by one focused build. We offer a free consultation to scope that first step, and we'll tell you if the honest answer is to wait or use a subscription tool instead.