Skip to main content

AI Consultant Questions: The 20 They Should Ask You Before They Quote

Most founders go into a first meeting with a list of AI consultant questions to ask. Fair enough. The more useful test runs the other way, though. Pay attention to the questions they ask you. A good one spends the first hour asking about your business, your team and the things that only live in […]

A person holds a tablet with a completed checklist while asking an AI consultant questions at a desk.

Most founders go into a first meeting with a list of AI consultant questions to ask. Fair enough. The more useful test runs the other way, though. Pay attention to the questions they ask you. A good one spends the first hour asking about your business, your team and the things that only live in your head. A bad one spends it showing you a demo.

I’ve sat on both sides of that table. I’ve been the business owner getting pitched tools that didn’t fit, and these days I spend my weeks doing discovery with founders. Below are the twenty questions I think any AI consultant worth hiring should ask you before they quote a thing. They’re grouped into four areas, with red flags at the end. If they skip most of them, walk.

Key Takeaways

  • The AI consultant questions asked in the first meeting tell you more about their quality than any demo, case study or slide deck they show you.
  • A good AI consultant asks about your business, your team and your bottlenecks before they mention a single tool, platform or software subscription.
  • If they never ask which decisions only you can make, they cannot fix the real problem: a business that runs out of the founder’s head.
  • Consultants who ask first will map your data sources and current numbers, so any result they promise can be measured against a real starting point.
  • The best first automation is usually boring, recurring admin that eats hours every week, not the complex judgement calls you hoped AI would take.
  • Every proposal should say what a human must always own, who looks after the system after launch, and how your customer data will be protected.
  • If a consultant cannot tell you what success looks like at 30 and 90 days, you are buying a project with no finish line. Keep looking.

The AI Consultant Questions About Your Business (1 to 5)

The research on AI projects is brutal. MIT’s 2025 work on generative AI in business found around 95% of pilots delivered no measurable return. The ones that worked had something in common: they started with the business problem, not the technology.

That’s why the first five questions should have nothing to do with AI at all. If someone opens with their favourite platform, they’ve already told you how the project will go.

1. What does a good week look like in your business?

This sounds soft. It isn’t. A consultant needs to know what “working well” means to you before they can spot what’s getting in the way. For a trades business, it might be every quote out within 24 hours. For a clinic, it might be a full book and no chasing. Your answer becomes the target everything else gets measured against.

2. If you got 15 hours back each week, what would you do with them?

This one separates consultants who sell efficiency from the ones who sell outcomes. Freed time with no plan for it just fills back up with firefighting. A good consultant wants to know whether you’d spend those hours on sales, on a new service line, on hiring well, or on getting home before 7 pm. That answer shapes what gets automated first.

3. Where does your revenue actually come from?

Not the pitch deck version. The real one. Which services make the money, which clients are worth the most, and which channels bring them in. Automating something that touches 5% of your revenue is a nice-to-have. Automating something in the path of your best clients is where the return sits.

4. What would break if you disappeared for two weeks?

Every founder knows the answer to this instantly, and it’s usually uncomfortable. Quotes stall. Staff wait. Invoices don’t go out. Clients ring your mobile. This question shows a consultant where the business depends on one person, which is almost always the real constraint. If they don’t ask it, they’re probably about to automate around the edges instead of fixing the middle.

5. What have you already tried, and why did it stall?

Most business owners I talk to have tried something. A CRM nobody updates. A chatbot that annoyed customers. A ChatGPT subscription that helped for a week. A good consultant wants the post-mortem, because the reasons things stalled (no adoption, no context, no owner) are exactly what the next build has to solve. If you’ve been burned before, I’ve written about starting AI automation again after a bad vendor.

Three colleagues in an office, one addressing AI consultant questions on the phone while working at a desk, while the others review documents.

Questions About Your Team and What Lives in Your Head (6 to 10)

Here’s the pattern I see in nearly every founder-led business between 5 and 50 people. The team is capable. They just don’t have the context to make decisions without the owner. So they ask. All day. And the owner answers, because it’s faster than explaining the whole picture.

An AI consultant who doesn’t dig into this is going to build you tools. A good one is going to build the business a brain.

6. Which decisions can only you make?

Get them to list them with you. Pricing a non-standard job. Whether to take on a difficult client. Which lead to chase first. When founders actually write these down, the list is always longer than they expect, and a surprising number of items turn out to be decisions the team could make if the reasoning were captured somewhere. That reasoning is what an AI brain holds.

7. What does your team ask you every day?

This is the gold mine. The repeat questions (“Where’s the latest price list?”, “How do we handle this client?”, “Can I offer a discount here?”) are a map of the knowledge that hasn’t left your head yet. Each one is a candidate for the system to answer instead of you. A consultant who asks for examples, not a summary, is doing the job properly.

8. What’s written down, and does anyone actually use it?

Most businesses have a folder of SOPs somewhere. Most of those SOPs haven’t been opened since the week they were written. A good consultant asks this because the answer tells them whether you have a documentation problem or an access problem. Usually it’s access. A document sits there. A system that has read the document and answers questions from it gets used.

9. How long does a new hire take to become useful?

If the answer is three to six months, that’s a sign the business knowledge lives in people, not systems. Every new starter has to absorb it by osmosis, mostly from you. This question also links straight to cost. A slow ramp means you’re paying a salary for months before you see the output. A consultant who connects onboarding time to the knowledge problem understands what they’re looking at.

10. Who on your team would own this day-to-day?

Building the system is maybe 30% of the work. Getting people to use it is the other 70%. If no one on your team is asked to champion it, test it and feed back on it, the build becomes another tab nobody opens. A consultant who asks this early is planning for adoption, not just delivery. One who never asks is planning to hand over a login and leave.

AI Consultant Questions About Your Numbers and Tools (11 to 15)

You can’t improve what you can’t see. Before anyone promises a result, they need to know where you are now. These questions build the baseline, and they also reveal the quick wins that tend to pay for everything else.

11. What tools do you use, and do any of them talk to each other?

Accounting, CRM, booking system, project management, email platform, a spreadsheet someone updates on Fridays. The typical founder-led business runs on five to eight systems that don’t share data. A good consultant lists them all with you. Not to rip them out, but to work out how to connect them so the AI can see the whole picture instead of fragments.

12. Which numbers do you check every morning, and where?

If you’re logging into four dashboards before breakfast to work out how yesterday went, that’s a problem a system can solve. This question tells the consultant which three to five numbers actually matter to you. It’s also the starting point for a daily brief: one summary, pulled automatically from your real data, waiting on your phone when you wake up.

13. How fast do you respond to a new enquiry?

Research from InsideSales and Harvard Business Review found that the first business to respond wins the deal around 78% of the time. Most small businesses take hours, sometimes days. An AI consultant should want your real number, not your hopeful one. If it’s slow, that’s often the single fastest return in the whole engagement, because the leads are already arriving and you’ve already paid for them.

14. What’s sitting in your customer database?

Old quotes that never closed. Past clients who haven’t heard from you in two years. Enquiries that went cold because nobody followed up. Most businesses have years of contacts they’ve never properly worked. A consultant who asks how many records you have, how old they are and what a customer is worth to you is looking for money you’ve already earned the right to ask for.

15. What happens to a phone call nobody answers?

This question gets awkward fast, which is the point. Plenty of businesses don’t know how many calls they miss. One dental client I work with, Dr Claire, had nearly half of calls going unanswered despite two receptionists at the desk. Once an AI receptionist was picking up overflow and after-hours calls, missed calls dropped to zero and booked appointments rose 44%. You can’t fix that if nobody asks. More on this in my post on after-hours call answering.

Two colleagues discuss AI consultant questions while reviewing a document and a laptop showing an upward-trending line graph in an office.

Questions About Risk, Ownership and Results (16 to 20)

This is the group most consultants skip, and it’s the group that decides whether you’re happy in six months. Anyone can build a demo. The hard parts are boundaries, privacy, ongoing care and an honest definition of success.

16. What must a human always handle?

There are things in your business that should never be handed to a machine. Delivering bad news to a client. Pricing a complex job. Anything where a mistake carries legal or reputational weight. A good consultant asks you to draw these lines before they build, and designs the system so it hands off to a person at exactly those points. If they think everything can be automated, they haven’t run a business.

17. What customer data will this touch, and where can it live?

Your clients trust you with their information. Any AI build will touch some of it. A consultant should ask what data is involved, who can see it, and whether any of it carries obligations under New Zealand’s Privacy Act or your industry’s rules. The Office of the Privacy Commissioner publishes clear guidance on this. If your consultant has never mentioned privacy, raise it yourself and watch how they answer.

18. Who looks after this once it’s live?

AI systems aren’t set and forget. Your business changes. Prices change. Staff change. The models change every few months. A good consultant asks how you want the system cared for after launch: whether your team will maintain it, or whether you’d rather have ongoing training, optimisation and support from them. The worst answer is no answer, because that’s how builds quietly rot.

19. What does success look like at 30 and 90 days?

Not “it’ll save you time.” Specific outcomes, agreed in writing. Response time under five minutes. Fifteen hours a week back from admin. A set number of reactivated customers. A good consultant pushes you to name these before work starts, so you both know whether it’s working. I’ve written a whole piece on running a proper 90-day AI automation checkpoint if you want a template.

20. What would make you say this failed?

This is the question I respect most when someone asks it, because it takes nerve. It forces both of you to be honest about expectations up front. Maybe failure means your team won’t use it. Maybe it means customers complain. Maybe it means you’re still the bottleneck in month three. Naming it early means you can design against it, rather than discovering it later.

Red Flags: When to Walk

You’ve now got twenty questions to listen for. You won’t hear all twenty in a first meeting, and that’s fine. Discovery takes time. But if most of them never come up, here are the signs you’re talking to a tool seller, not a partner.

They lead with the platform. If the first ten minutes are about their favourite software, their model or their agent builder, they’ve decided the answer before hearing the problem. Tools are the last decision, not the first.

They quote before discovery. A fixed price for “AI automation” after one 20-minute call means they’re selling a package and fitting you into it. A good consultant wants to understand the business first, even if that means a paid diagnostic before a full proposal.

They promise percentages with no baseline. “We’ll save you 40% of your admin time” means nothing if nobody has measured your admin time. Questions 11 to 15 exist for a reason.

They can’t explain it in plain English. If you walk out confused about what they’ll build and how it helps, that confusion will follow you through the whole project.

Nobody mentions what happens after launch. If question 18 never comes up, you’re buying a build with no care plan. Six months later it breaks, and nobody owns it.

It works in reverse too. You should be asking them hard questions as well, about contracts, ownership and exit terms. My post on the questions your accountant would ask about an AI automation contract is a good starting list. And if you’re still weighing whether AI is right for your size of business at all, start with AI adoption for SMEs.

A man in a cafe leans over his laptop to video chat with a woman wearing a headset to answer AI consultant questions.

Conclusion

The best AI consultant questions are about your business, not about AI. What a good week looks like. Which decisions only you can make. What your team asks you every day. Where the numbers live, how fast you respond, and what success looks like in 90 days. A consultant who asks those things is trying to understand the problem before selling you a solution.

The pattern underneath all twenty is the same. Most founder-led businesses don’t have a tools problem. They have a business that runs out of one person’s head. The fix is an AI brain that holds what you know, and an AI workforce that handles the recurring work, so the business stops waiting on you. Any consultant worth hiring will see that within the first hour, because they asked.

Want Someone to Ask You These Questions?

If you’d like to see what this looks like in practice, book a 30-minute discovery call. I’ll ask you a good chunk of the questions above, and you’ll leave with a clearer picture of where your time goes and what’s worth fixing first. No demo, no pitch deck. Just a conversation about your business.

Frequently Asked Questions

What questions should I ask an AI consultant before hiring them?

Ask how they run discovery, what they need from you, and how they measure success. Ask who owns the system and the data once it’s built, what ongoing support looks like, and what happens if you part ways. Ask for an example of a project that didn’t go to plan and what they learned. Their answers matter less than whether they ask about your business first.

How much does an AI consultant cost in New Zealand?

It varies widely. Some charge hourly for advice only, some sell fixed packages, and others run a paid diagnostic before scoping a build. The cheapest option is rarely the best value if it skips discovery. Look for a clear scope, agreed outcomes and a plan for support after launch. Compare what you’ll get back in time and revenue, not just the upfront figure.

What does an AI consultant actually do for a small business?

A good one maps how your business runs, finds where time and money leak, and works out which recurring tasks a system can take over. Then they build or recommend the setup, connect your existing tools, train your team and keep improving it. The goal is fewer things routing through the owner, faster response to customers and better visibility of the numbers.

How do I know if my business is ready for AI?

You’re likely ready if you’ve been trading for a couple of years, have at least one staff member, and feel like the business stops when you step away. Other signs include slow lead follow-up, missed calls, a database you never work and too many dashboards. You don’t need perfect systems first. A good consultant starts with whatever you already have.

How long does it take to see results from AI automation?

Quick wins such as faster lead response, an AI receptionist or reactivating an old database can show results within weeks. Bigger changes, like capturing what lives in the founder’s head and building a daily brief, take a few months to bed in properly. Agree on 30-day and 90-day targets before work starts so you can judge progress on evidence, not on feel.

Should I hire an AI consultant or just use ChatGPT myself?

ChatGPT is useful for one-off tasks, but every conversation starts from scratch, and it knows nothing about your business. A consultant builds a system that holds your context, connects to your data and runs tasks without you prompting it. If you have the time and interest, DIY can work. Most busy founders find the time cost is the very problem they’re trying to solve.

About Octavius

Titus Mulquiney is the founder of Octavius AI, where he builds AI brains and AI workforces for founder-led businesses stuck running everything out of their own head. Twenty years in marketing, ex-Sony product manager, ex-GM Zeal NZ. Based in Auckland, working with operators across NZ, Australia, and the US. Connect on LinkedIn.

The Business Stress Test

Stuck running your business out of your head? Three minutes on your phone tells you what the chasing costs you a year, where it leaks, and what to hand off first.

Brass Butler
Brass tacks, deftly handled.

Strategic automation for founders who want their time back. Starts with a free 30-minute call to find out if there's a fit.

Book a discovery call

Not ready for a call? Take the three-minute Business Stress Test.