The twelve questions that separate AI automation consulting firms that deliver measured results from ones that sell software — with the answers you should expect to hear.
Every consultant sounds the same on a discovery call: proven process, tailored solutions, cutting-edge AI. The differences only show up in the specifics — so make the specifics the interview. Here are the twelve questions we'd ask before hiring any AI automation consulting firm, including the answers that should make you comfortable and the ones that should end the meeting.
The twelve questions
- What's your process before you propose anything? Good answer: a workflow audit of our actual business. Bad answer: a demo of the platform they always recommend.
- Will you quote a fixed price for a defined first project? Good answer: yes, after the audit. Bad answer: 'we work on a monthly retainer' before scope exists.
- Who owns the automations when we part ways? Good answer: you do — your accounts, your data, documented handover. Bad answer: anything involving their proprietary platform.
- Which tools do you build on, and why those? Good answer: whatever you already use, plus mainstream automation tools. Bad answer: a single vendor they happen to resell.
- What happens when the AI gets something wrong? Good answer: a human-review step for anything below a confidence bar, plus logging. Bad answer: 'it doesn't.'
- What data leaves our systems, and where does it go? A serious firm can name the AI providers involved, what's sent, and how the setup respects PIPEDA.
- How will we measure success? Good answer: a before/after baseline in hours or dollars, checked at 30 and 90 days. Bad answer: 'you'll feel the difference.'
- Can we see a comparable project? Ask for one at your size — a 200-person deployment says nothing about a 6-person office.
- Who actually does the work? The person selling isn't always the person building. Meet the builder.
- What won't you automate? Firms with judgment have a ready list (complaints, judgment calls, anything reputation-critical). Firms without one automate whatever you'll pay for.
- What does maintenance look like? Automations touch other people's software; things drift. Good answer: a light support arrangement with response times, not a surprise invoice.
- What happens in month two if the automation isn't saving what you projected? The answer tells you who carries the risk — listen carefully.
Red flags that override good answers
- Pressure to sign a long retainer before any audit has happened.
- Case studies with percentages but no baseline ('40% more efficient' — than what?).
- No written scope — 'we'll figure it out as we go' means the budget will too.
- Everything is possible. Real builders say 'that part's hard' about something.
You're not hiring intelligence — artificial or otherwise. You're hiring accountability for a measured result.
We publish our own answers to all twelve questions in how our AI automation consulting engagements are structured — audit first, fixed scopes, your accounts, measured outcomes. Bring this list to a call with us and ask them in order; we think interviews like that are where good firms win.
