How to Choose an AI Consultant: Start With the Problem.

Most owners who go looking for an AI consultant don't actually want AI. They want the phone to ring more, the leads they already get to stop slipping away, or the thing that slows the business down every week to finally get fixed. AI is one way to get there. The trouble is that "AI consultant" now covers everyone from strategy shops that never ship anything to agencies reselling someone else's software under their own logo. The way through is to name your problem first, then hire the kind of help built to solve it.

Which problem are you actually trying to solve?

In service businesses doing $1M to $10M a year, growth almost always gets stuck in one of three places. Each one needs a different kind of help.

1. Not enough leads. More of your customers now ask ChatGPT or Google's AI who to hire, and those assistants name two or three businesses. If you aren't one of them, you never see the lead you lost. What you need is someone who checks where you stand, fixes what the assistants read about you, and reports whether it is working.
2. Leads that go cold. The calls that go to voicemail after hours, the web form nobody answers until Monday, the quote that never gets a second follow-up. What you need is a system that answers every inquiry within minutes and keeps following up, plus someone who runs and tunes it every month after launch.
3. A bottleneck inside the business. The handoff that stalls, the report rebuilt by hand, the process that lives in one person's head. What you need is someone senior who finds the real bottleneck, ranks the fixes by return, and builds them, without you hiring a full-time executive.

If you can't tell which one is yours, that is normal. Answering it is the first job of any consultant worth hiring, and the ones who skip straight to a product demo are telling you something.

What are the four business models hiding behind one title?

1. The strategy shop. Produces a roadmap, a maturity assessment, and a slide deck. Genuinely useful if your problem is executive alignment. Useless if your problem is that nothing is in production. The tell: they cannot name the systems their last three clients now run.
2. The reseller agency. Configures a third-party platform and bills it as a custom build. Fine for commodity workflows, and often the cheapest correct answer — but you own nothing, and your run cost is whatever the platform decides it is next year. The tell: the "custom" system cannot be exported or migrated.
3. The contract dev shop. Builds exactly the spec you hand them. Excellent when you have a competent internal technical owner writing that spec. Dangerous when you do not, because they will faithfully build the wrong system.
4. The build-and-run firm. Decides what to build, builds it, and stays accountable for how it performs every month after launch. The only model where the consultant carries the consequences of a bad decision, and the only one that fits problems like lead follow-up, where the system has to be tuned as your business changes.

None of these is fraudulent. The failure mode is buying one and believing you bought another. Strategy alone won't answer a single extra lead, and a reseller won't find a bottleneck their product doesn't touch. For the closely related question of firm structure, see AI Development Firm vs. AI Agency.

What questions expose which model you are talking to?

Ask these in the first conversation. They are not gotchas — a good partner answers all of them without hesitation, and the hesitation itself is the signal.

• "What will you measure, and what do I get if you miss it?" Leads, response time, booking rate, hours saved. A firm confident in its work names the number and puts a guarantee behind it. Vague answers here predict vague results.
• "Who reviews the system after launch, and how often?" A named role and a stated cadence, or the answer is nobody.
• "Show me a log of what the AI said to a real customer." Redacted is fine. A firm that runs systems for clients has this. A firm that ships demos won't know what you are asking for.
• "What do I own at the end?" Code, AI instructions, call and message logs, and your customer data, listed in writing. "You own your data" is a non-answer; everyone says that.
• "What is the monthly run cost, and which parts of it are yours versus passed through?" Model inference, telephony, hosting, and their margin should be separable line items.
• "Which project of yours failed, and why?"Anyone who has shipped enough AI systems has one. A firm with no failures has no production history, or is not being candid.

How do you judge a consultant without being technical?

You do not need to assess model architecture. You need to assess whether the firm makes decisions deliberately. Three proxies work reliably:

• Do they push back on your brief? A firm that agrees with everything in the first call is selling, not designing. The most valuable early meetings we have are the ones where we tell an owner that what they asked for is the third thing they should fix, not the first.
• Do they distinguish between what the system decides and what it recommends? This distinction is the core of responsible AI design, and firms that have not thought about it will not have language for it. The govern and manage functions of the NIST AI Risk Management Framework exist precisely because that boundary is where AI systems cause harm.
• Do they talk about your numbers before their tools? Research on where organizations capture value from AI — including ongoing work published by MIT Sloan Management Review — consistently points at organizational and process factors rather than model selection as the differentiator. A firm that leads with which AI model it uses is answering the least important question. One that asks how fast you answer leads today, or where jobs get stuck, is starting in the right place.

What should the commercial terms actually say?

Most disputes we hear about are not about quality. They are about terms nobody read carefully. Five clauses are worth negotiating properly:

1. Ownership and escrow. Enumerate the artifacts. If the firm hosts the system, specify what you receive if the relationship ends — ideally a working export, not a zip file of source you cannot deploy.
2. Run-cost transparency. Pass-through costs (inference, telephony, hosting) itemized separately from the management fee, with a stated policy for what happens when usage exceeds the assumed volume.
3. Change control. Who can change the system's behavior, and does that change get logged? An AI system whose instructions can be edited silently by anyone with dashboard access is unauditable regardless of what the contract says.
4. Performance definition. The number that matters, how it is measured, and what happens if it is missed, in the contract, not in an email.
5. Exit. Notice period, transition support, and the state the system is left in. A partner confident in the work makes this easy.

When is a consultant the wrong answer entirely?

Sometimes the honest recommendation is not to hire one. In our engagements, the cases where outside help is the wrong call cluster tightly:

• An off-the-shelf product already fits. If a product you can buy does most of what you need, buy the product. Custom work is for the part of your business that is actually different. We lay out the test in Custom AI vs. Off-the-Shelf.
• The data is not ready. If the system of record is contested — two CRMs, three price lists — no AI system will resolve that, and building one on top of the ambiguity just makes the ambiguity faster.
• Nobody internally will own it. Without a named human in command, a custom system degrades quietly. Hire that person first.
• You already have the team. If you have engineers with production AI experience and the capacity to use them, the trade-off is different — we lay it out in Strategic AI Consultants vs. In-House AI Team.

The short version

Name the problem first: not enough leads, leads going cold, or a bottleneck inside the business. Then hire the kind of help built for that problem, from a firm that will put a number on the outcome, show you real logs, list what you own, and stay accountable after launch. Filtering takes one conversation.

If you're not sure which problem is yours, the free two-minute assessment sorts it out, and our services page explains how we handle each one. For the failure patterns that make this choice matter, see Why AI Projects Fail at $1M–$10M Businesses.