How to Hire an AI Consultant (Without Getting Scammed)

There's a gold rush happening in AI consulting right now. Everyone with a LinkedIn Premium account and a ChatGPT subscription is suddenly an "AI transformation strategist." They've got the slide decks. They've got the buzzwords. They've got testimonials from companies you've never heard of. What they don't have is a single system running in production.

This isn't a small problem. Small businesses are spending $10,000-$50,000 on AI consulting engagements that deliver nothing but a PDF of recommendations and a handshake. The consultant moves on to the next client. The business owner is left with a strategy document nobody knows how to execute and a healthy skepticism about whether AI was ever going to work for them.

It was going to work. They just hired the wrong person. Here's how to make sure you don't make the same mistake.

The Red Flags: Walk Away If You See These

1. They Promise "10x ROI" Before Understanding Your Business

This is the biggest red flag in AI consulting, and it's everywhere. You hop on a discovery call. Within the first ten minutes — before they've asked a single question about your operations, your data, or your problems — they're already telling you about the incredible returns you'll see.

"Our clients typically see 10x ROI within the first year." "AI can reduce your operating costs by 40%." "Companies like yours are leaving millions on the table."

These aren't insights. They're sales scripts. Nobody can tell you what AI will do for your business before they understand your business. A consultant who leads with ROI projections before doing any discovery work is telling you they sell the same pitch to every prospect. They're not solving your problem — they're fitting your problem into their solution.

A legitimate consultant's first call should be 80% questions and 20% talking. They should be asking about your workflows, your pain points, your team, your data, and your goals. The ROI conversation comes after they understand what they're working with — not before.

2. They Can't Show You Something They've Built

Ask every AI consultant this question: "Can you show me a working system you've built for a client?" Not a slide deck. Not a mockup. Not a case study written in the third person with no verifiable details. A living, functioning thing that exists in the world and does something useful.

If the answer involves a lot of NDAs, vague references to "enterprise clients," or a pivot to talking about their "methodology," you're talking to someone who advises on AI but doesn't build it. Those are two very different skill sets, and only one of them produces results.

The AI consulting market is full of people who understand the technology conceptually but have never shipped anything. They can tell you what's possible. They can draw architecture diagrams. They can run workshops. But when it comes time to actually build, configure, deploy, and maintain a system that works in your environment with your data — they subcontract it, delay it, or scope-creep it into oblivion.

The question isn't "do you understand AI?" The question is "have you built something with AI that's running right now, in production, for a real business?" If the answer is no, keep looking.

3. They Want to Sell You a Platform Instead of Building Custom

Be very cautious of consultants who show up to the discovery call with a product already picked out. "We're partnered with [Platform X] and we'll set you up on their system." That's not consulting. That's reselling.

There's nothing inherently wrong with platforms. Some are excellent. But a consultant whose recommendation is predetermined isn't evaluating your needs — they're steering you toward whatever pays them the highest commission or affiliate fee. The platform might be perfect for you. It might also be wildly overbuilt, underbuilt, or completely wrong for your use case. You'll never know, because the "evaluation" was a formality.

A good consultant starts with your problem and works backward to the solution. Sometimes that's a platform. Sometimes it's custom-built. Sometimes it's a $20/month tool you didn't know existed. The answer should depend on what you need, not on what the consultant sells.

4. The Timeline Is Vague or Absurdly Long

"We'll start with a 3-month discovery phase, followed by a 6-month implementation roadmap, with full deployment expected in 12-18 months."

For a small business? No. Just no. If an AI consultant can't deliver something useful — not perfect, not complete, but useful — within 4-8 weeks, either the scope is wrong or they're padding the engagement. Enterprise AI projects take 12 months. Small business AI projects that take 12 months are small business AI projects that never finish.

The right approach for most small businesses is a focused pilot: one problem, one solution, 4-6 weeks, measurable results. If it works, you expand. If it doesn't, you've learned something valuable for a fraction of the cost. Anyone pushing a multi-phase, multi-quarter engagement before proving the concept on a single use case is optimizing for their revenue, not your results.

5. They Can't Explain What They'll Build in Plain English

If you ask "what exactly will you build?" and the answer includes the phrases "neural network architecture," "proprietary algorithm," "deep learning pipeline," or "enterprise-grade AI framework" — and you still don't understand what the thing does — that's a problem.

Good consultants explain things simply because they understand them deeply. The jargon isn't a sign of expertise. It's a smokescreen. The best AI solutions for small businesses aren't technically exotic. They're well-built systems that connect the right AI capabilities to specific business workflows. If the consultant can't explain that in a sentence, they either don't know what they're building or they don't want you to know.

The Right Questions to Ask

Now that you know the red flags, here's your actual playbook for evaluating an AI consultant. These five questions will separate the real ones from the noise.

"Show me something you've built that's running in production right now."

This is the single most important question. Production means live, in use, by real people, doing real work. Not a demo environment. Not a proof of concept from 2024. Something that's running today. If they can show you a system — even if it's for a different industry — that's handling real data and producing real results, you're talking to a builder. If they can't, you're talking to an advisor.

"What's your timeline from kickoff to first working deliverable?"

The answer should be in weeks, not months. Four to six weeks for a focused pilot is reasonable. Eight weeks for something more complex is understandable. Anything beyond that for a first deliverable means the scope is too big or the consultant is too slow. Either way, it's a problem.

"What happens if it doesn't work?"

This is where you learn someone's character. A consultant who's never had a project go sideways is either lying or hasn't done enough projects. Things fail. Data is messier than expected. The use case that seemed obvious turns out to be harder than anyone thought. How the consultant handles that matters more than their success rate.

Good answers: "We scope the pilot to limit your exposure." "We've had projects where we pivoted to a different approach after the first two weeks." "Here's an example of something that didn't work and what we did instead."

Bad answers: "That won't happen." "Our methodology guarantees results." "We haven't had a failed project." (Translation: they redefine success to match whatever happened.)

"What will I own when the engagement ends?"

This is critical and often overlooked. Some consultants build on proprietary platforms that you can't access without them. Some retain ownership of the code. Some create dependencies that ensure you need them forever — which is great for their recurring revenue and terrible for your independence.

You should own everything: the code, the data, the infrastructure, and the knowledge to maintain it. A good consultant builds you something and teaches your team how to run it. A bad consultant builds you something that only they can operate.

"Can you explain your approach in terms my team will understand?"

The best AI implementations succeed not because of the technology but because the team adopts them. If the consultant can't explain what they're building, why it matters, and how it will change daily workflows in language your operations manager or office staff can follow, adoption will fail. Technical brilliance that nobody uses is the most expensive kind of waste.

What a Good Engagement Actually Looks Like

For reference, here's what a well-run AI consulting engagement looks like for a small business. Not every engagement follows this exact pattern, but the principles are consistent:

1. Discovery (1-2 weeks). The consultant interviews your team, observes workflows, audits your data, and identifies the highest-impact opportunity. This is where they earn their fee — by finding the right problem to solve.
2. Scoping and agreement. A clear, written scope: what will be built, what data it needs, what the expected outcome is, what the timeline is, and what success looks like in measurable terms. No ambiguity.
3. Build and iterate (3-5 weeks). The consultant builds the solution, tests it with real data, and iterates based on results. You should see working progress weekly — not monthly status reports, but actual demos of things that work.
4. Deployment and training (1 week). The system goes live. Your team is trained. Documentation is provided. You know who to call if something breaks.
5. Measure and evaluate (2-4 weeks post-launch). Compare before and after. Hours saved, errors reduced, costs eliminated, revenue gained. Real numbers, not estimates.

Total timeline: 6-10 weeks from first call to measurable results. Total investment for most small businesses: $5,000-$25,000 depending on complexity. For context, here's our detailed breakdown of how much AI consulting actually costs.

If someone is quoting you six months and six figures for your first AI project as a 20-person company, they're scoping an enterprise engagement for a small business budget. Walk away.

The Market Is Getting Better (Slowly)

The good news: the AI consulting market is starting to self-correct. The first wave of hype merchants is being exposed as businesses compare notes and realize that slide decks don't produce ROI. The consultants who survive will be the ones with production systems, real case studies, and clients who actually got results.

The bad news: we're still in the messy middle. For every legitimate AI consultant, there are five people who rebranded from "digital transformation" or "blockchain consulting" and are selling the same empty playbook with new terminology. Your job as a buyer is to tell the difference. The questions above will get you there.

If you want to understand more about what a legitimate consultant actually does day-to-day, our guide on what an AI consultant actually does breaks it down in detail.