Skip to content
CentriWeb
Not sure where to start?

Tell us how the business runs. We'll tell you what we'd build first.

Book a call
For builders and tradesPreBuildMore homeowner enquiries, more trust, more jobs won. For sole tradersOwner CFOA financial operating system for owners who do the books.
Need something of your own?

We build custom software too.

Talk to us

AI consultant for small business: what they do and how to choose one

An AI consultant for a small business advises, builds, or builds and looks after. Here is how to tell which you need, what a first engagement should produce and what to ask.

An AI consultant for a small business finds the jobs in your business where AI or automation would save time, bring in more money or stop you losing money, and then either writes that plan down, builds it, or builds it and looks after it. The first thing to settle is which of those three you are paying for, because the price, the risk and the result are different for each.

What an AI consultant for small business actually does

The title covers very different work. One person sells a workshop on using AI chat tools. Another reviews your processes and hands you a report. Another connects your phones, website, CRM and accounting so that jobs happen without anyone remembering to do them. All of them are called consultants.

The useful work has three parts: working out where your time or money is going, choosing the smallest change that fixes it, and making sure the change keeps working after the first week. A consultant who skips the first part sells what they already have. One who skips the last part leaves you with something that breaks quietly.

Plenty of businesses do not need one. If your problem is a single repeated task, such as writing first drafts of emails, an AI assistant and an afternoon may cover it. Our guide to where to start with AI shows how to try that yourself first.

Three kinds of help, and what each one means for you

Kind of help What you receive Who does the work Suits you if
Advice only A written plan, priorities and a list of tools You or your staff You have someone with time and technical interest to carry it out
Build The system set up and handed over The consultant builds; you run it You can maintain it, or the system is simple and stable
Build and look after The system, plus fixes and changes after launch The consultant, on an ongoing basis Nobody on your team wants to own it, or it touches customers every day

The gap that catches people is the middle one. A build with no care afterwards works until a phone system changes, a form is edited or a tool alters how it connects. Then nobody knows who fixes it. Ask who answers when it breaks, and whether that is included in the price or billed separately.

What a first engagement should produce

Whatever the kind, the first stage should leave you with things you can read and check:

  • A list of your costliest problems, each with a rough figure for time or money, using numbers from your own records.
  • One job chosen to start with, and the reason it comes first.
  • Which of your existing systems will be connected, and what will stay as it is.
  • What a person still does, and what happens when the system is unsure.
  • How success will be measured, the starting figure, and the date you will check it.
  • Who looks after it afterwards, and what that costs.

If the first deliverable is a general presentation about AI trends, you have bought education, not a plan for your business.

Red flags

  • They propose tools before asking about your business. A fair proposal starts with your calls, enquiries, jobs and paperwork.
  • Promised results with no baseline. Nobody can tell you how much time you will save without seeing how you work today.
  • Guaranteed outcomes or timeframes. Too many things outside their control affect the result.
  • Everything is custom, or nothing is. A good adviser says when an off-the-shelf tool is enough and when it is not.
  • You would not own it. Ask who holds the accounts, the logins and the data, and what happens if you leave.
  • Vague answers on customer data. They should say where your customers’ details go and which AI services see them.
  • No one says what a person still does. Systems that answer customers need a handover to a human.

A simple test applies to all of these: if a claim cannot be checked against your own records after the work is done, do not pay for it as if it could.

Questions to ask before you sign

  1. Which of the three kinds of help is this, and what is excluded?
  2. What will you need from me and my staff, and how many hours is that?
  3. Which of my systems will you connect, and what happens if one of them changes?
  4. Who owns the accounts, the logins and the work if we stop?
  5. Where does my customer information go, and which AI services see it?
  6. What does it cost to run each month, including tool subscriptions and usage fees?
  7. Can you describe a similar job you did, what went wrong and how you fixed it?
  8. What will we measure, and when do we review it?
  9. What would make you tell me not to do this?

The last question matters most. A good consultant has an answer. For a longer checklist, see our guide on how to choose an AI automation agency.

What to prepare before you talk to anyone

You will get a better answer, and a fairer price, if you arrive with facts. Spend an hour on this first:

  • Write down the three tasks that eat most of your week, and roughly how many hours each takes.
  • List the systems you already pay for: phone, website, CRM, accounting, booking and email.
  • Count how many calls and enquiries you miss or answer late in a typical week.
  • Decide who in your business would own the result, and how much of their time they can give.

With those four things, you can compare proposals on the same terms. Without them, every consultant fills the gaps with their own assumptions. If you are still deciding whether AI is worth the effort at all, our guide on whether AI is worth it for a small business works through that first.

Judging value: time saved, money made or money kept

AI work earns its fee in one of three ways. It saves time you value. It brings in money you would not have had. Or it stops you losing money you already make. Anything outside those three can wait.

Take an example with invented numbers. A small plumbing business misses ten calls a week while the owner is on site, and one in five of those would have booked a job worth $400. That is two jobs, or $800 a week, which is a good reason to look at call answering. If the same business spends five hours a week typing up invoices and chasing payments at a loaded cost of $40 an hour, that is $200 a week of time. Put your own figures in, write down the real counts, and compare the yearly total with the cost of the build and the monthly care.

If the saving is smaller than the cost, say so and walk away. Our guide to measuring the return on AI shows how to set the baseline.

Where CentriWeb fits

CentriWeb is a founder-led technology studio based in New Zealand that works online with businesses worldwide. We build and look after systems such as AI phone answering and lead follow-up, starting with the part that costs the business most, and we say so when an off-the-shelf tool is enough. We are the wrong choice if you only want a workshop or a report, or if one person on your team has the time and interest to set up the tool themselves.

Common questions.

What does an AI consultant do for a small business?

They find the jobs in your business where AI or automation would save time, bring in money or stop you losing it. Some stop at a written plan, some build it, and some build it and look after it. Ask which one you are buying before you sign.

How do I know if I need an AI consultant at all?

You may not. If your problem is one repeated task and an existing tool already covers it, set it up yourself or ask your software provider. A consultant earns their fee when the job crosses several systems, such as phones, your CRM and your accounting, or when nobody on your team has time to run it.

What should a first engagement produce?

A written list of your costliest problems with a rough time or money figure against each, one job chosen to start with, and a clear statement of what success looks like and when you will check. It should name which systems of yours will be connected and who will look after the work afterwards.

How do I judge whether the work paid for itself?

Pick one of three measures before the work starts: hours saved per week, extra money brought in, or money you stopped losing. Record the starting number from your own records, then compare after a fixed period. If nobody will agree to a measure, treat that as a warning.

Related guides.

AI for business · 5 min read

How to automate tasks with AI in a small business

Talk to CentriWeb about this

CentriWeb helps businesses put AI to work where it saves time or money, and keeps it away from the jobs it should not do. The first call takes 30 minutes and is free.

Or call +61 483 957 775 or email info@centriweb.com. CentriWeb is based in New Zealand and works online with businesses worldwide.

Ask a questionBook a call