The 10 questions to ask an AI consultant.
Ask them in this order. The first half tells you whether they can do the work. The second half protects you after you sign.
1. Which workflow would you start with, and what number would change?
A good consultant names one workflow your team runs every week and the number it should move, such as hours spent on weekly reporting or how long a new lead waits for a reply. They also explain how they will measure the starting point before anything gets built. Be careful with an answer that opens with a company-wide AI strategy. A first engagement should prove value on one piece of real work, and the roadmap can grow from there.
2. What will it cost, and is the price in writing before work starts?
Ask for a fixed price tied to a written scope. Open-ended hourly billing puts the risk of a slow project on you. Published small business AI audit prices run $1,500 to $5,000 (see what an AI audit costs), and this price breakdown covers monthly advisory and builds. Ask about running costs after launch too, like model usage and software seats, so the first-year number holds up.
3. Who will actually do the work?
Plenty of firms sell with a senior person and deliver with someone you never meet. Ask who leads your sessions and who you call when something breaks. Then ask to talk with that person before you sign. On a small engagement, the person on the sales call should be the person giving you advice.
4. Can I see something you built that is running today?
Ask for a screen share of a working system, with client details hidden. Then ask what went wrong while they built it. People who have shipped real work answer that one quickly and with specifics. Case study PDFs with no names and no demo are hard to check, so give them less weight.
5. Who owns what you build, and where does it run?
Your prompts and workflows should live in accounts your company controls, with documentation your team can read. If the system only runs inside the consultant's own platform, you are renting your own process, and leaving gets expensive. Ask what a handover looks like on the last day of the project.
6. Where does our data go, and who can see it?
Ask which AI providers will touch your data and what their data terms say. Ask what access the consultant needs, and push for the least access that gets the job done. A first conversation should need no access at all. If someone asks for an admin password on the first call, end the call.
7. Which steps stay with a person?
Anything that touches money or customers needs a named person approving it, at least at the start. A good answer points to the exact step where the system stops for review and says who reviews it. If the pitch says AI will handle everything on its own, keep looking.
8. When would you tell us not to use AI?
This is the most revealing question on the list. Some workflows run too rarely to be worth automating, and some depend on judgment that should stay human. You want to hear specific examples of when they told a client to wait. A consultant who has never said no to a client is selling you something.
9. Which AI tools do you work with, and are you tied to one?
Most small teams already pay for ChatGPT or Claude and use it for a fraction of what it can do. Ask whether the consultant will build on what you already have before suggesting new licenses. Partner status with a vendor is useful, because it brings access to that vendor's programs and support. It should never stop them from recommending a different tool when it fits better.
10. How do we leave if it is not working?
Ask about the contract length and the notice period. Month-to-month advisory with 30 days notice keeps a consultant earning their place every month. For builds, ask what your team needs to run the system without the consultant, and get that in the scope.
Write each answer down. Two consultants who sound alike on a call usually look very different on paper.
Ask us all ten, free
The free AI assessment is a 30-minute call with Mitchell about one real AI decision. Bring this list and ask every question on it. You leave with your next actions in priority order, whether we work together or not.
The call needs no access to your systems or files. If none of our offers fit, we say so.
“They are now doing 2 to 3x the work they were doing before.”Engineering team leadAfter our workshops
How OperatorZero answers the 10 questions.
We hold ourselves to the same list. Every answer below comes from our published pages.
| Question | Our answer |
|---|---|
| Which workflow first? | One workflow your team runs every week. The $2,500 AI workflow audit prices every step and ranks the 3 to 5 places AI pays for itself. We fix the top win during the audit. |
| What does it cost? | Every price is published, with no hourly billing. Advisory starts at $2,500 a month. Builds start at $6,000, at a fixed price agreed in a written scope. |
| Who does the work? | Mitchell Moses leads your sessions and answers your questions himself. Our team handles builds and workshops. |
| Something running today? | A real estate firm's files turned into a searchable knowledge base with Claude Code, with people reviewing the output. Read the case study. |
| Where does it run? | Builds run inside the tools your team already uses, and we train the people who use them. |
| Our data? | The free call needs no access to your systems or files. Reviewing real data is paid audit work, scoped before it starts. |
| What stays human? | Agent workflows stop for review, and a person approves what matters. |
| When do you say no? | If the audit finds nothing worth building, the report says so with the numbers. Sometimes the right next step is to wait. |
| Tied to one vendor? | We are a Claude partner and an OpenAI select partner, and we still tell you when another tool fits better. |
| How do we leave? | Advisory runs month to month with 30 days notice. The audit fee is credited toward a build or your first month of advisory. |
Red flags worth walking away from.
- A price you only see after a paid discovery phase.
- Hourly billing with no cap and no written scope.
- Case studies you can't check, and no live demo.
- A request for admin passwords or shared logins.
- Guaranteed savings before anyone has timed your work.
- A system that only runs on the consultant's platform, with no way to export it.
Advisory or an audit first?
Once you trust a consultant, match the engagement to the problem. Need ongoing direction and a second opinion on every AI pitch? That is monthly AI advisory, from $2,500 a month, with written next steps after every session.
Want hard numbers on one workflow before you commit? Start with the $2,500 workflow audit. The fee is credited toward a fixed-price build from $6,000 or your first month of advisory, so you don't pay twice.
Consultant or a full-time hire?
A full-time AI engineer runs about $142,000 in base salary before benefits. Until AI work fills a full-time role, a monthly advisor costs less and starts this month. When you outgrow advisory, we will help you hire that person.
Want to price a workflow yourself first? The free DIY audit checklist walks you through it.
Questions.
Anything else, bring it to the free AI assessment call.
How do I choose an AI consultant for a small business?
Ask each candidate to name the first workflow they would measure and to put a fixed price in writing. Then ask when they would advise against AI, and how you would leave if it isn't working. Pick the one with the most specific answers.
What questions should I ask an AI consultant before hiring?
Ask which workflow they would change first and what it costs in writing. Then ask who does the work, whether you can see a system running today, who owns what they build, where your data goes, which steps stay with a person, when they would say no to AI, whether they are tied to one vendor and how you leave.
How much does an AI consultant cost for a small business?
Published small business AI audit prices run $1,500 to $5,000. At OperatorZero, monthly advisory starts at $2,500 a month and builds start at $6,000. Every price is fixed and published. See every price on one page.
Should I hire an AI consultant or a full-time AI person?
A full-time AI engineer runs about $142,000 in base salary before benefits. Until AI work fills a full-time role, a consultant on a monthly plan usually costs less and starts sooner. See monthly AI advisory.
What are the red flags when hiring an AI consultant?
A price you only see after a paid discovery phase is one. A request for admin passwords on the first call is another.
Do we need technical people on our team?
No. A good consultant explains everything in plain language and matches the advice to the team you have.
Is the free AI assessment really free?
Yes. It is a 30-minute call with no obligation, and it needs no access to your systems or files. You keep the next actions whether we work together or not. See what the call covers.
Want a straight answer on AI for your business?
Bring these ten questions to a free 30-minute call with Mitchell. See what the free AI assessment covers.