What does an AI consultant actually do?
Before asking questions, it helps to be clear on what you're buying. AI consultancy broadly covers three things: assessment (understanding where AI can help your business), advisory (recommending what to do and in what order), and implementation (actually making it happen).
Some consultants do all three. Many specialise in one or two. Knowing which of these you need — and which the consultant you're speaking to actually delivers — should be the starting point for any conversation.
Questions worth asking
Have you worked with businesses like ours? Not in terms of size or sector necessarily, but in terms of where they are on the AI adoption curve. Ask for specific examples of similar engagements and what the outcome was.
What do you actually deliver, and what do I walk away with? Some consultants deliver a report. Others deliver a working implementation. Be specific about what you expect to have at the end of the engagement and make sure there's alignment before you start.
How do you approach the discovery phase? Any consultant worth hiring will want to understand your business before recommending anything. Ask how they do that — do they come on-site, speak to multiple people across the business, or work primarily from documents? The depth of the discovery phase is usually a reliable indicator of the quality of what follows.
What happens if the recommendation is to use existing tools rather than build something custom? A good AI consultant gives you honest advice about what your business actually needs, even if that means recommending a £30/month SaaS tool rather than a bespoke build. If a consultant seems more interested in selling a development project than solving your problem, that's worth noting.
Who does the work? In larger consultancies, the person who sells the engagement is often not the person who delivers it. In smaller or solo practices, you work directly with the consultant throughout. Neither model is inherently better, but you should know which one you're signing up for.
What does success look like, and how will we measure it? If a consultant can't answer this question clearly, that's a problem. Good consultancy produces measurable outcomes — time saved, cost reduced, capability built.
Red flags to watch for
A consultant who leads with technology rather than problems. The question should always start with what challenges the business faces, not which AI tools are currently popular.
Guaranteed outcomes. AI implementations involve uncertainty. Anyone promising specific results before they've understood your business is overselling.
No interest in your existing tools or processes. Recommendations that ignore what's already in place are often impractical to implement and miss quick wins that are already available.
Pricing that's entirely opaque. You should understand what you're paying for before you commit. Scoped, fixed-price engagements are generally preferable to open-ended day rates for defined pieces of work.
What good looks like
A good AI consultant listens more than they talk in the early stages. They ask questions about the business before making recommendations. They're honest about what AI can and can't do. They give you a clear view of what working together looks like, what it costs, and what you'll have at the end.
And they're as comfortable recommending a simple, inexpensive solution as a complex one — because the goal is to help your business, not to maximise the size of the engagement.
Hiring an AI consultant is a significant decision. The questions above won't guarantee a perfect outcome, but they'll give you a much clearer picture of who you're dealing with — and whether they're the right fit for what your business actually needs.