How to Choose an AI Consultant

Why This Guide Exists

Yes, we are an AI consultancy. Yes, we are writing a guide on how to choose an AI consultant. We understand the irony. But here is why we are doing it anyway.

The AI consultancy market is flooded with providers who oversell, underdeliver, and leave business owners worse off than they started. We hear about it constantly — businesses that spent thousands on a "digital transformation" and ended up with a half-built system nobody uses. Or who paid for a "strategy" that was a generic slide deck with their logo pasted on top.

We would rather you chose the right consultant — even if that is not us — than chose badly and concluded that AI does not work. Bad consultants do not just waste money; they poison the well. They make business owners cynical about AI, which means they miss the genuine opportunities.

So here is an honest guide. Use it to evaluate anyone, including us. If we do not meet these standards, tell us.

What to Look For

Not all AI consultants are created equal. Here are the qualities that separate the good ones from the ones who just talk a good game:

Industry experience that is relevant to yours. An AI consultant who has only worked with tech startups will struggle with the realities of a construction business or a hospitality group. Look for someone who understands your industry's specific challenges, regulations, and workflows. They do not need to be a specialist in your exact niche, but they should demonstrate genuine understanding of your world.

A structured methodology. Good consultants have a clear, repeatable process for how they work. They can explain exactly what happens at each stage, what you will receive, and what they need from you. If someone cannot explain their methodology clearly, they are probably making it up as they go.

Willingness to say no. This is the single best indicator of a trustworthy consultant. If someone tells you AI is the answer to every problem without understanding your business first, they are selling, not consulting. The best consultants will tell you when AI is not the right solution, when you need to fix foundations first, or when a simpler approach would work better.

References you can actually check. Ask for references from businesses similar to yours. Then actually call them. Ask what went well, what did not, and whether they would hire the consultant again. The answers will tell you more than any case study or testimonial.

UK-based and accessible. Time zones matter. Understanding UK business culture, regulations, and market conditions matters. Being able to visit your premises matters. Remote work is fine for much of the project, but having a consultant who can sit with your team when needed is valuable.

Red Flags to Avoid

These warning signs should make you seriously reconsider a potential consultant. Any one of them is concerning. Two or more means walk away.

Questions to Ask

Here are ten specific questions to ask any AI consultant you are considering. Each one is designed to reveal something important about how they work. Ask them all, and pay close attention to how confidently and specifically they answer.

1. "Can you walk me through a similar project you have completed?" — You want specifics, not vague generalities. What was the business? What was the challenge? What did they build? What were the results?
2. "What does your discovery process look like?" — A good consultant will describe a structured approach to understanding your business before recommending anything. Look for process mapping, staff shadowing, and data assessment.
3. "When would you tell a client NOT to use AI?" — The answer reveals honesty. If they struggle to think of a scenario, that is a red flag.
4. "How do you measure success?" — Look for specific, measurable metrics defined before the project starts, not vague promises of "improvement."
5. "What will we own at the end of the project?" — You should own everything: the automations, the documentation, the data. If they retain ownership or lock you into their platform, think very carefully.
6. "Who specifically will work on our project?" — In larger consultancies, the person who sells you is often not the person who does the work. Know who you are actually getting.
7. "What does post-implementation support look like?" — Automation needs maintenance. What happens when something breaks at 9pm on a Tuesday? What is included and what costs extra?
8. "What do you need from us?" — A good consultant will be clear about the time, access, and resources they need from your team. Vague answers here lead to unexpected demands later.
9. "Can we speak to a reference who had a problem during their project?" — Every project has problems. What matters is how they were handled. A consultant who only offers glowing references is hiding something.
10. "What happens if the project does not deliver the expected results?" — Listen for honesty and a plan, not defensiveness or unrealistic guarantees.

What a Good Engagement Looks Like

A professional AI consultancy engagement follows a predictable structure. If a consultant proposes something radically different, ask why. Here is what you should expect:

Discovery (1–2 weeks). The consultant learns your business. They review your processes, talk to your team, assess your data, and understand your goals. This should feel like a thorough investigation, not a quick scan. If you have already completed our AI Readiness Checklist, share the results — a good consultant will appreciate the preparation.

Shadowing and documentation (1–2 weeks). The consultant observes how your team actually works — not how the manual says they should work, but what really happens. This is where the real insights emerge. Processes get documented, bottlenecks get identified, and automation opportunities get mapped.

Recommendation and planning (1 week). You receive a clear recommendation: what to automate, in what order, at what cost, with what expected returns. A good recommendation includes what NOT to automate, and explains why.

Implementation (4–8 weeks for a first project). The automation gets built, tested, and refined. Your team is involved throughout — providing feedback, testing, and learning. This is not something that happens in isolation and gets dropped on your desk.

Handover and support (ongoing). Your team is trained to manage the automation day-to-day. Documentation is complete. Support arrangements are clear. You should feel confident running the system without the consultant on speed dial.

How to Compare Proposals

When you have proposals from two or three consultants, compare them on these dimensions rather than just price:

Scope clarity. Does the proposal clearly define what is included and what is not? Vague scoping leads to scope creep, budget overruns, and disappointment. The more specific the proposal, the better.

Timeline realism. Be wary of proposals that promise to transform your business in two weeks. A realistic first automation project takes 6–12 weeks from discovery to handover. Shorter timelines usually mean corners are being cut.

Pricing transparency. Can you clearly see what you are paying for? Are there likely additional costs? What happens if the project takes longer than expected? The best proposals include a breakdown of costs by phase and clear terms for changes.

Support terms. What happens after go-live? How long is support included? What does ongoing maintenance cost? This is where the real long-term cost lives, and it is the part most people forget to compare.

Do not automatically choose the cheapest option. The cheapest consultant often becomes the most expensive when the project overruns, deliverables are substandard, and you need to hire someone else to fix the mess. Equally, the most expensive option is not automatically the best. Value is the metric that matters.

Making Your Decision

After reviewing proposals, checking references, and asking your questions, the decision often comes down to trust. Do you trust this consultant to be honest with you? Do they understand your business? Did they listen more than they talked? Did they ask smart questions?

Here is a simple framework: choose the consultant who asked you the best questions. The one who wanted to deeply understand your business before recommending solutions. The one who was honest about limitations. The one whose methodology made you think "that makes sense."

We are one option among many. We welcome you to use this guide to evaluate us alongside anyone else. If you would like to start that conversation, get in touch — our initial consultation is free, and we will be honest about whether we are the right fit.

Here are the key takeaways:

• Look for industry experience, structured methodology, and willingness to say no
• Run from guaranteed ROI promises and tool-pushing before discovery
• Ask all ten questions and check references — actually call them
• Expect a structured engagement: discovery, shadowing, recommendation, implementation, support
• Compare on scope clarity, timeline realism, and pricing transparency, not just cost
• Trust your instinct — choose the consultant who asked the best questions