Demand for AI consulting is growing faster than the supply of quality. Alongside experienced practitioners, the market is filling up with tool vendors who have six months of experience and a confident voice. For a company, telling them apart is hard — both talk fluently about agents, automation and savings. These 7 questions will show you the difference.
1. Do they ask about your processes first, or go straight to a tool?
This is the most reliable tell. Someone who opens with “we'll roll out ChatGPT/Copilot for you” is selling a tool. Someone who opens with questions — where time gets lost in your company, what people do by hand, where decisions are made by gut feel — is consulting. An AI project without a process audit is a lottery ticket.
2. Have they ever run a company or a team?
Introducing AI is a change project, not a technical installation. It fails on people, unclear briefs and missing support from leadership — exactly where every other change in a company fails. A consultant who has never led a team doesn't see these obstacles until they hit them. With you.
3. What will the output look like?
Ask for a concrete answer: a document with priorities and expected return, a trained team, a working automation. If the answer is a “workshop” with no defined output, you are paying for time spent, not for a result.
4. How do they measure success?
Hours saved per week, how fast enquiries get handled, how many people are actually using AI a month later. Someone who can't tell you what will be measured and when isn't planning to come back to the result.
5. Do they train on generic examples or on your tasks?
Generic “what AI can do” training gets people excited and wears off within a month. Training on your people's real tasks — your invoices, enquiries, reports — gets used the next day. Ask what exactly participants will be working on.
6. Are they independent of the tools they recommend?
Commission on licences doesn't have to be a problem — but you should know about it. An independent consultant can also say “you don't need AI for this, a better process will do”. That very sentence is the best sign of quality.
7. What happens after the project ends?
AI tools change from month to month. A one-off project without follow-up support means your workflows are out of date within six months. Ask about follow-up: consultations after rollout, updates, availability for questions.
Red flags
- They promise specific savings before they have seen your processes.
- They sell with a grant from a programme that no longer runs (DIGI pro firmu, a Czech digitalisation grant, ended in November 2025 — current state of grants here).
- They can't explain what they do without buzzwords.
- They don't have a single real example from an industry similar to yours.
- The price is still “depends on scope” after the second meeting. Everyone should have an indicative price list — here's mine.