Twelve years in AI and automation · Auckland, New Zealand 021 365 082 Taking new projects Book a free call
Home What I build Case studies FAQ Newsletter Contact Book a free call

Guide · Choosing an AI consultant

How to Check an AI Consultant in New Zealand Is Legit Before You Hire Them

Anyone can use the title. Here is what to check in ten minutes before you talk to them, what to ask on the first call, and the signs that should make you walk away.

Updated6 October 2026
Reading time8 minutes
BasedAuckland, working NZ-wide

Why you have to check

AI consulting is not a licensed profession in New Zealand, so the title tells you nothing. A lot of people added "AI" to their LinkedIn headline in 2023. Some of them are good. You cannot tell which from the headline.

Visitors to my own website ask the assistant on it the right questions, usually bluntly: "how many clients", "who owns ez ai", "do u have qualification". This page is how to get real answers to those, about me or about anyone else you are considering.

Ten minutes of checking before you talk to anyone

Most of what you need is public, and you can check it before you book a call.

  • The Companies Office and NZBN registers. Search the business name. It should be a registered New Zealand company with a named director, and you can see how long it has existed. Both searches are free.
  • GST registration. Ask for it if it is not on their site. A consultancy doing real volumes of business work will be registered, and their invoice should carry the number.
  • Google reviews. Look for named people describing a specific job. Five reviews that say what was built beat fifty that say "great service".
  • The LinkedIn profile of the person who will do the work. What were they doing before 2023? Years of building software, running projects or working with data is a good sign. A career that began with ChatGPT is a risk.
  • Their own website. If they sell chatbots, is there one on their site, and is it any good? If they sell automation, does their own booking and follow-up run on it? People who build this for a living use it themselves.
  • Case studies with numbers. A sector, a process, hours or dollars saved, and how long it took. "We transformed their operations" is not a case study.

Eight questions for the first call

The first call should be free, and these eight questions will tell you most of what you need. The right-hand column matters as much as the middle one.

AskA good answer sounds likeBe wary if
Can you show me something you built that is still running?A screen share of a live system, or a client who will take your call.You only ever see slides and demos.
Have you automated a process like mine?A similar process, even in another industry, and what went wrong along the way.Everything went perfectly, every time.
What will it cost, and when will I know?A fixed price in writing before any work starts.An hourly rate and "we'll see how it goes".
Who owns the code and the accounts?You do. It runs in accounts in your name and is documented so someone else could take over.It lives on their platform and stops if you leave.
Where will my data be, and does anything train a public model?A named country and cloud provider, and a plain no on training."It's in the cloud" or "it's secure".
What happens when the AI gets something wrong?It runs in test mode first, flags what it is unsure of, and a person signs off anything that matters. Every decision is logged."It doesn't make mistakes", or no mention of testing.
Who will actually do the work?A name, and you can talk to them.The person selling is never the person building.
What happens if you disappear?The code, the documentation and the accounts are already yours, so another developer could pick it up.The question seems to offend them.

One more that is worth asking: "When would you tell me not to use AI?" A lot of the work I am asked to do with AI is better done with plain rules-based automation, which is cheaper and more reliable. Someone who never says that is selling a tool.

"Do you vibe code or actually code?"

Someone typed that into the chat on my site, and it is a fair question. Almost every developer now writes code with AI help. I do, every day. What matters is whether the person can read what the AI wrote, test it, and fix it when it breaks.

So ask how the thing will be tested, where it will run, and who gets the alert when it fails. If the answers are specific, the code is in good hands however it was typed. If the answer is that the AI handles all that, it will be your problem within a few months.

The same goes for no-code tools. Zapier and Make are the right answer for simple jobs, and a good consultant will say so. I have compared them here: Make vs Zapier vs a custom build.

Red flags

Any one of these is a reason to slow down. Two is a reason to walk away.

  • Paid discovery. A competent consultant can scope a first project in a free conversation.
  • Hourly billing on a build, with no cap. All the risk sits with you. Hourly is fine for small fixes and advice. A build should have a price.
  • A proprietary platform you cannot leave. Ask what you would be left with if you stopped paying.
  • Guaranteed returns. Nobody can promise a figure before measuring your process.
  • Straight to live. No test mode, no trial run against real data.
  • AI for everything. If every problem you describe needs an agent, they are not listening.
  • Pressure to sign this week. The work will still be worth doing next month.

Solo consultant or agency?

Both can work. They fail in different ways. With a solo consultant you talk to the person doing the work and you are not paying for account managers. The risk is capacity: one person can be sick, on holiday or busy. With an agency there are more hands, but there may be two or three people between you and whoever is actually building it.

I am a solo consultant, so here is the question to put to people like me: what happens when you are away? A good answer covers monitoring that raises an alert without anyone watching, documentation another developer could follow, and accounts in your name. I also do not promise 24/7 support, because one person cannot honestly offer it.

How I answer these

Briefly, because you should check them, not take my word for it. EZ-AI.NZ LIMITED is a registered, GST-registered New Zealand company. I have spent twelve years in AI and automation as an analyst, product manager and engineer. Every figure on this site comes from a real project. Builds are fixed price, you own the code and the accounts, hosting is AWS Sydney in your own account by default, nothing trains a public model, and you deal with me from the first call to support.

The detail is on the AI consultant page, the reviews are on Google, and the work is in the case studies.

Frequently asked questions

Is AI consulting regulated in New Zealand?

No. There is no licence, registration or required qualification for calling yourself an AI consultant in New Zealand. The usual laws still apply, including the Fair Trading Act, the Consumer Guarantees Act where relevant and the Privacy Act 2020, but nobody vets the title. That is why checking shipped work and references matters more than checking certificates.

What qualifications should an AI consultant have?

There is no standard one. Vendor certificates show someone passed a course on a product. What predicts a good result is a track record: systems they built that are still running, in businesses willing to say so. Ask how long they have been doing this kind of work and what they were doing before 2023.

How many clients should an AI consultant have had?

Enough that they can show you two or three finished projects with real numbers, and at least one client who will take a call from you. A long logo list matters less than one system that has been running for a year.

Should an AI consultant have worked in my industry?

It helps but it is not essential. The process matters more than the sector: quoting, intake, claims, scheduling and reporting look much the same everywhere. Ask for the closest process they have automated, not the closest industry.

What should a first AI project cost?

For a New Zealand small or medium business, most first projects land between NZ$500 and $15,000 as a fixed price agreed before work starts. Be wary of a first project quoted in the tens of thousands, or of any build billed by the hour with no cap.

Put these questions to me

A free 30-minute call is a fair test. Ask me anything on this page, bring the process that is costing you time, and judge the answers for yourself. Discovery is always free.

Book a free call