How to Vet an AI Consultant Before You Write the Check
The Market Is Flooded
When ChatGPT went mainstream, something predictable happened: everyone became an AI consultant overnight.
Agencies rebranded. Freelancers updated their LinkedIn headlines. Former SaaS reps added "AI strategy" to their service offerings. There's no certification required, no body of work required, no barrier to entry at all. Anyone can print new business cards and start charging for it.
That's not a criticism of the industry. It's just the reality you're operating in right now. You are trying to make a smart hiring decision in a market that makes that very hard to do.
If you're a founder or operations leader at a service business, you're probably fielding outreach from people who can talk about AI for 45 minutes without saying anything specific. You've seen the decks. You've heard the promises. And you're right to be skeptical.
But there are tells. The good ones and the bad ones both show up fast — if you know what to look for.
What a Real AI Consultant Actually Does
The word "strategy" does a lot of work in consulting pitches. It sounds like a deliverable. It isn't.
Real AI integration consulting produces things you can point to. Documented workflows. Configured automations that are live and running. Measurable before-and-after data on time spent, error rates, or throughput. A consultant who does real work for small businesses leaves behind systems your team can actually use — not a deck they present and never touch again.
The difference comes down to build versus advise. An advisor tells you what's possible and hands you a roadmap. A consultant who's done business AI implementation builds the thing, trains your people, and is accountable to an outcome. One of these is dramatically more valuable for a 30-person service business.
Here's what a legitimate first engagement actually looks like in practice: they audit your existing processes to find where time is being lost or work is falling through the cracks, they scope a focused pilot on one workflow — not your whole operation — and they define what success looks like before anyone writes a line of code or buys a subscription. Six weeks later, you either hit the target or you have a documented reason why you didn't.
This isn't about finding someone smart. Plenty of smart people have no business being inside your operations. It's about finding someone who's done this for businesses like yours — same size, same constraints, same kind of work.
That's a much shorter list, and they're worth finding.
Five Questions to Ask Before You Hire Anyone
Don't wing the intro call. Go in with these:
Can you show me a before/after from a client in a similar business? Not a case study with vague percentages — a real walkthrough of what the problem was, what was built, and what changed. If they can't produce this, that tells you something.
What does the first 90 days look like, specifically? You want a week-by-week answer, not a phases diagram. Anyone who's actually done this before can tell you exactly what happens when.
Who does the actual implementation — you, or a subcontractor? There's nothing wrong with using specialists, but you need to know who you're actually working with and who's accountable when something goes sideways.
How do we measure success, and what happens if we don't hit it? A good consultant will have already thought about this. If they're surprised by the question, that's a red flag.
What do you need from my team, and how much of their time? Workflow automation for small business only works if your people are involved. A consultant who can't answer this hasn't thought through the implementation at all.
Red Flags That Should End the Conversation
Some of these will show up on the first call. Pay attention.
Jargon without substance. There's nothing wrong with technical language when it's being used precisely. But if someone is talking about LLM pipelines, agentic workflows, and RAG architectures without being able to tell you in plain English what problem that solves for your business — they're either performing expertise or they don't know your industry well enough to be useful.
Vague case studies. "Helped a client increase efficiency by 40%" is not a case study. 40% of what? Measured how? Over what period? Real results are specific. If they can't be specific, assume there's a reason.
Tool-first thinking. If a consultant's first move is to sell you on a particular platform before they've understood your operation, they're working backwards. The tool should follow the process diagnosis, not the other way around. This is how you end up with AI tool graveyards — subscriptions piling up, half of them unused.
Retainer-only pricing with no milestones. A monthly retainer with no defined deliverables is a cash flow problem wearing a consulting hat. Tie every engagement to outcomes.
Can't tell you what they won't do. This one matters more than people think. A consultant who knows their lane — and is honest about the edges of it — is far more trustworthy than one who says yes to everything. If they can't tell you what's outside their scope, they either haven't thought about it or they're not being straight with you.
What a Good First Engagement Looks Like
If you find someone worth hiring, here's what the engagement structure should look like.
Scoped, not open-ended. A fixed project with clear outcomes. Not "ongoing AI strategy support" — a specific problem, a specific solution, a specific result.
Starts with a process audit. Before anyone recommends a tool or writes an automation, they need to understand what your team actually does all day. This is where fixing the process before you automate it matters — automating a broken workflow just makes the broken thing happen faster.
Deliverables you own. Documented SOPs. Configured systems. Trained staff. When the engagement ends, you should have assets — not dependencies. A slide deck you can't execute without the consultant is not a deliverable.
Accountability built in. Weekly check-ins. Agreed metrics established at the start. A clear exit plan so the handoff doesn't evaporate.
This is how Recursive Solutions structures every engagement. Before any work starts, we run a growth mapping call to identify where automation will actually move the needle for your specific operation — not where it looks impressive in a deck. A lot of what we find is operational, not technical. That's usually where the real leverage is. It also helps avoid one of the most common failure modes in this space — diving into AI for service businesses without first understanding what the business actually needs, which is one of the main reasons AI projects fail at the SMB level.
Ready to Cut Through the Noise?
If you've been burned before, or you're trying to figure out who to trust before you write a check, book a free 30-minute growth mapping call through the Map Your Growth contact form on the Recursive Solutions site.
Worst case, you walk away with a clear picture of where automation can actually help your business — insight your competitors are paying consultants good money for. Best case, you find a partner who's done this before and can show you the work.
Either way, it's 30 minutes. Go find out.