AI for Small Business: Where to Start With One Job 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 for small business: where to start

Where to start with AI in a small business: pick the one job that costs the most hours, trial it for two weeks, and keep a person checking the output.

The best place to start with AI for small business is the single job that costs your team the most hours and follows the same steps each time. Pick that one job, trial an AI tool on it for two weeks, and keep a person checking everything it produces before a customer sees it.

Most owners who feel burned by AI started with a tool and went looking for a use. It works better the other way around. The tool is the cheap part. The job, the information it needs and the person who checks it are what decide whether it helps.

Where AI for small business pays off first

AI is worth paying for when it does one of three things: it saves time, it brings in more money, or it keeps money you already make. If a proposed use does none of those, it can wait.

  • Saves time: drafting replies, summarising long emails, writing up meeting notes, pulling figures off invoices.
  • Brings in money: answering new enquiries quickly, including after hours, and following up quotes that have gone quiet.
  • Keeps money you already make: reminding clients about appointments, chasing overdue invoices, and recalling past customers who are due for a service.

Look at your own week for work that is repetitive, rule-based and low risk if it goes slightly wrong the first time. That is where AI is most reliable.

How to find the one job

Ask the people doing the work, not the person with the most opinions about AI. A short exercise:

  1. List every task your team repeats weekly. Include the dull ones, such as copying details between systems.
  2. Next to each, write roughly how many hours it takes.
  3. Circle the ones that follow clear steps, where you could explain the right answer to a new starter in ten minutes.
  4. Cross out anything where one wrong answer would be costly: legal wording, medical advice, tax figures, safety decisions.
  5. From what is left, take the biggest.

That is your first project. Write it down in one sentence, for example: “Answer the common questions that arrive by email and text, and hand anything unusual to Sam.” If you run a business in New Zealand, our guide to AI for small business in NZ adds the local points to check.

Examples from ordinary businesses

A plumbing business. The office manager spends the first hour each morning replying to overnight enquiries. Most ask the same things: do you service my suburb, can you come today, what do you charge for a call-out. An AI assistant can draft replies from the business’s own service area and price list. The office manager reads each draft, fixes anything wrong and sends it. A quick check takes much less time than writing from scratch.

An accounting practice. Staff chase clients for missing documents every month. An AI tool can read an incoming email, work out which document it is, and file it against the client. A person still checks anything the tool was unsure about, and checks the figures before they reach a return.

A physio clinic. Patients who have not rebooked after treatment drift away. A system that texts them a booking link after a set number of weeks keeps those patients without anyone remembering to do it. A staff member decides the wording and who gets contacted.

A builder. Quotes go out and nothing happens for a fortnight. Drafting a polite follow-up for each open quote is time the owner does not have. The owner reads each message before it goes. There is more on this in our guide on how builders follow up quotes.

What a person must still check

AI tools produce confident text that is sometimes wrong. They can invent a price, misread a handwritten figure, or promise something your business does not offer. The risk is not that the tool fails loudly. It is that it fails quietly in a sentence that reads well.

Whatever you choose, someone on your team should still check:

  • Any price, date, availability or promise in a message going to a customer.
  • Any figure taken from a document before it goes into your accounts.
  • Anything involving a complaint, a refund, a safety matter or a health detail.
  • Summaries of long documents, where the tool may leave out the one clause that matters.

Our guide on reducing AI mistakes covers how to cut the number of wrong answers, and the checks above catch what is left.

Keep customer information safe from day one

Before anyone pastes customer names, phone numbers, health details or financial information into an AI tool, find out what the tool does with it. Some free tools may keep what you type; some business plans say they do not. Read the provider’s privacy terms, and if you cannot tell, keep personal details out.

In Australia, the Privacy Act applies to many businesses, and not every small business is covered, so check whether it applies to yours. The Office of the Australian Information Commissioner publishes guidance on using commercial AI products on oaic.gov.au, and it is worth reading before you choose a tool. Our guide on using ChatGPT safely in a small business lists practical rules for staff, and a one-page AI policy puts them in writing.

Questions to ask before you pay for anything

If a vendor, or anyone selling you AI, wants money before the trial, these questions separate a tool that will help from a demo that looks good:

  • Which of my jobs does it do, and which does it not do? A vague answer is a bad sign.
  • Where does it get its answers? A tool that answers from your own price list and service area is safer than one that guesses from the internet.
  • What happens when it is unsure? It should pass the job to a person, not carry on.
  • What does it do with the information my customers give it, and where is that stored?
  • Can I leave? Ask what you keep if you stop paying: your data, your phone number, your automations.
  • Who looks after it when something changes, such as a new price or a new service?

Write down the answers. If you decide not to go ahead, you will not have to rely on memory a month later.

A two-week trial that tells you something

Do not roll out to the whole business. Run a small trial:

  • Before: time the job for a normal week. Write down the hours and the number of mistakes you find.
  • Week one: one person uses the tool on the job, with everything checked before it goes out.
  • Week two: same again. Note every time the tool got something wrong, and why.
  • After: compare the hours, including checking time, and the mistakes.

If the job is clearly faster and the error rate is no worse, keep it and think about the next job. If you spent as long fixing as you would have spent doing, drop it without guilt. Our guide on measuring the return on AI goes further.

What can wait

You do not need an AI strategy document, a dedicated hire, or a tool for every department. Skip anything that is mainly impressive in a demo. If a use does not save time, bring in money or keep money, leave it for later.

Some jobs are worth having set up properly rather than trialled by hand, such as an assistant that answers from your own information, books into your real calendar, writes to your CRM and hands over to a person when it should. That is the kind of AI work CentriWeb builds, and it makes sense once you know which job is worth it. For a small business with a few enquiries a week, an off-the-shelf tool and a careful person are often enough.

Common questions.

What is the first thing a small business should use AI for?

The job that costs the most hours and follows the same steps every time, such as answering the same enquiries, drafting routine emails, or chasing paperwork. Pick one, not five, so you can tell whether it worked.

Do I need to buy special software to start?

Not always. A general AI assistant you already have access to can draft emails, summarise documents and tidy notes. Software built for one job only becomes worth paying for once you have a task that repeats often enough to justify it.

Is it safe to put customer details into an AI tool?

Check the tool's privacy terms first and keep personal and financial details out of any tool you have not checked. Write a short rule for staff, and keep a person responsible for anything that goes to a customer.

How do I know if AI is saving me money?

Time the job before you start, then time it again after two weeks, including the checking. If the job is faster and the quality is the same, it is working. If you spend as long fixing the output as you did doing the job, stop.

Related guides.

AI for business · 6 min read

AI hallucinations: why AI makes things up and how to reduce them

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.

CentriWeb is based in New Zealand and works online with businesses worldwide.

Want this working in your business? We'll map it with you on a call.

A 30 minute call. You'll leave knowing what we'd build first.

Book a call