AI consulting prices range from $200/hour to $1,000/hour for nominally the same work. Quality variance is even wider. Use this checklist before signing any AI consulting engagement — the 12 questions that separate operators who deliver from advisors who just present.
AI consulting in 2026 is where digital marketing consulting was in 2010: a real and valuable category, with a flood of mediocre practitioners attracted by the demand. The variance between top-quartile and bottom-quartile AI consultants is much wider than for established categories — and the wrong choice wastes 6 months and $20K+.
The same words ("AI strategy," "implementation," "deployment") mean very different things across consultants. The checklist below tests for substance under the language.
1. Can they show you a working Claude Project they've built? The actual artifact — system prompt, knowledge base, examples. Not a screenshot. If they can't demo real implementations, they're an advisor, not a builder.
2. Can they walk through a specific past engagement in 10 minutes? What they did, what got built, what changed in the client's metrics. Specifics, not "we transformed their operating model."
3. Can they tell you the last 2 things that didn't work? Real operators have AI implementation failures in their history. Consultants who claim 100% success rates are presenting, not operating.
4. How do they prioritize what to deploy first? If the answer is "we do an audit," ask what the audit actually contains. If the audit is "we ask you what's painful and recommend AI for it," that's not methodology — that's order-taking.
5. What's their position on AI training cohorts vs. one-off workshops? Real operators believe in cohort training. Anyone who claims a one-day workshop produces adoption is selling something easy to deliver, not something that works.
6. Where do they think AI shouldn't be used in your business? Strong consultants have opinions on this. Weak ones say "AI can help with everything." The second answer means they haven't thought about it carefully.
7. Fixed price or hourly? Default to fixed price for any well-defined deliverable. Hourly aligns incentives against you. The exception: genuinely ambiguous discovery work that can't be scoped in advance.
8. Who actually does the work? Many consultancies sell on the senior person and deliver with juniors. Confirm who specifically will run your engagement.
9. What's the deliverable format? Slides are presentation artifacts. Operational deliverables are written docs, configured Claude Projects, runbooks. Strong consultants ship the operational stuff.
10. What happens after the engagement? Do you own everything that gets built? Can you operate it without them? If the answer is "well, you'll need our retainer to keep it running," you're being locked in.
11. What's the smallest engagement they'll take? Reflects market positioning. A consultancy whose minimum engagement is $50K isn't built for $5M businesses. Match scale to scale.
12. What does their roster of active clients look like? Mostly enterprise? Mostly SMB? Mostly your industry? Pattern matching matters — a consultancy with 5 active engagements in your category has compounding expertise you can't replicate by hiring.
Consultants who can't produce a written scope. Consultants who push for hourly billing on a well-defined deliverable. Consultants whose case studies are vague ("we helped a Fortune 500 company..."). Consultants who promise ROI numbers up front before doing diagnostic work. Consultants who haven't personally used Claude Projects (a tell that they're selling AI services without actually being AI operators).
Any of these is a hard stop. The downside of choosing wrong is much larger than the upside of choosing fast.
The right way to test fit before committing to a 90-day engagement: do a small productized audit first. $1,500–$5,000 for a 1-week diagnostic. You see how the consultant actually works, what the deliverable looks like, whether the recommendations are sound. Then decide whether to continue.
Treetop's $1,500 AI Audit is built for this exact test. 70% of audit clients continue into Implementation. The other 30% take the roadmap and execute themselves — which is fine.
— Bill Colbert, Treetop Growth Strategy