How to Choose an AI Consultant (From Someone Who Is One)

Every month, someone asks me: "We need AI advisory help. How do we pick a consultant?" The market is crowded with people calling themselves AI experts, and most of them are guessing.

This is uncomfortable to write because I have a financial interest—you hiring me instead of the wrong consultant is good for my business. But I'd rather you hire the right consultant (who might not be me) than hire the wrong one and waste six months.

Here's how to think about it.

Red Flags: Walk Away Immediately

1. They Promise a Fixed Outcome

"We guarantee 30% cost reduction," or "We'll deliver AI transformation in 12 weeks." This is a lie dressed as confidence.

AI adoption is dependent on your firm's data quality, technical infrastructure, team readiness, and specific problems. A responsible advisor tells you what's possible given your constraints, then works with you to remove constraints. They don't promise outcomes; they create conditions for outcomes.

Walk away from the guaranteed outcome pitch.

2. They're Selling a Specific Tool

If your consultant has a financial interest in you buying Tool A, they're not giving you unbiased advice. They might still be helpful—but the bias is real and material.

The right consultant recommends tools based on your problem. Different firms need different tools. If a consultant is evangelizing one platform, you should ask: "What does your firm gain if I buy this?"

3. They Don't Know Your Industry

AI in healthcare is different from AI in law is different from AI in engineering. If your consultant spent the first meeting asking basic questions about how your firm works, that's okay. If they spent 45 minutes pitching generic "AI strategy" without connecting to your specific challenges, they're not the right person.

Industry knowledge matters. Deep familiarity with professional services especially matters.

4. They Can't Explain What They Do in Plain English

If your consultant spends more than 30 seconds using jargon to describe their work, they're confused. Smart people can explain complex things simply.

Ask: "Walk me through a project you did last year. What was the problem? What did you recommend? What changed?" If the answer is clear and concrete, that's a good sign. If it's buzzword salad, move on.

Green Flags: This Person Might Actually Help

1. They Ask More Than They Tell

In your first conversation, a good consultant spends 60% of the time asking questions. "What's your current tech stack? Where are you losing time? What's your risk tolerance? How are your teams structured?" They're gathering data before forming opinions.

A bad consultant starts pitching in the first 10 minutes.

2. They Have Skin in the Game

The best consultants I know work on success-based fees or long-term retainers tied to realized outcomes. They have incentive to pick the right project and make sure it works.

You don't need to demand this model. But if a consultant suggests it, that's confidence.

3. They're Honest About What They Don't Know

"I haven't worked in aerospace, but here's what I'd research to understand your constraints"—that's credible. "I know exactly what aerospace needs"—that's a risk.

Look for consultants who are comfortable saying "That's a good question. Let me research that and come back with options."

4. They Have Real Implementation Experience

There's a difference between someone who reads about AI and someone who's actually built systems, managed integrations, and dealt with failure. Ask: "Tell me about a project that didn't go as planned. What went wrong? What did you learn?"

If they have no failure stories, they either haven't done much work or they're not being honest with you.

5. They Can Show Their Work

Case studies, project examples, client references (with permission)—these are credible signals. If a consultant can't or won't share concrete examples of work they've done, be skeptical.

The Interview Questions That Matter

When you're interviewing consultants, skip the pitch. Ask these:

• "What's one AI implementation you've done that failed? Why did it fail? What did you change in your approach after that?"
• "How do you measure success? What would success look like for my firm?"
• "Who on your team would actually work on this engagement? Can I talk to them?"
• "What's something I'm probably wrong about regarding AI and my industry?"
• "What would need to be true about my firm for this engagement to fail?"

The answers reveal a lot. Look for honesty, humility, and concrete thinking.

A Word About Cost

You get what you pay for, but price isn't the right variable. A bad consultant is expensive at any price. A good consultant is cheap at any reasonable price.

Compare consultants on: track record, relevant experience, team, and accountability mechanisms. Then compare cost. Not the other way around.

The Real Differentiator

At the end of the day, you're looking for someone who understands your business, can think clearly about AI, and gives a damn about whether it works. That person is harder to find than someone with a fancy credential.

Trust your instinct. If a consultant makes sense to you—if they ask the right questions and their answers land—go with them. If something feels off, it probably is.