The AI consulting market is crowded with two kinds of firms: those who ship working systems, and those who ship PowerPoints about working systems. These twelve questions — the ones we'd ask if we were hiring — reliably tell them apart in a single call.

Questions about proof

  • 1. Show me something you built that's in production today. Can I click it?
  • 2. What's a project that failed, and why?
  • 3. Which of your case studies can I verify with the client directly?

Questions about approach

  • 4. What would you NOT automate in our business? (Anyone who says 'everything's automatable' is selling.)
  • 5. Where do humans stay in the loop, and how do we measure when they can step out?
  • 6. Build vs. buy — when do you recommend off-the-shelf tools over custom builds?
  • 7. What happens to our data? Who trains on it, who stores it, where?

Questions about economics

  • 8. What's the smallest first engagement you offer? (Beware firms that only sell six-figure transformations.)
  • 9. How do you price — fixed scope, time and materials, or value-based? Why?
  • 10. What does maintenance cost after launch, honestly?

Questions about fit

  • 11. Who exactly will work on our account — the people on this call, or a bench we haven't met?
  • 12. If we should NOT do this project, will you tell us? Has that ever happened?
The best predictor of a good AI partner isn't their tech stack — it's whether they'll tell you when NOT to use AI.

For what it's worth, our answers to all twelve are on the table in every intro call — including the failed-project one. Book a free consultation and ask us anything on this list.