AI for small businesses: where it pays, and where it is a distraction

Every small business is being told it needs AI, and most of the advice starts with a list of tools. That is the wrong way round. The useful question is which part of your week AI could take on, and whether that part is what is actually holding the business back.

The answer usually splits three ways. Some jobs are a genuinely good fit for AI, and they are more ordinary than the demos suggest. Some uses are a distraction. And some problems look as if they need AI when they are really about how the business wins and keeps customers. That third group matters most.

Where AI pays

The pattern is the same in almost every business. AI earns its place where three things meet: work that repeats every day, a clear right answer or a first draft that is quick to check, and a person reviewing the result before it reaches a customer.

Sorting enquiries. Reading each new enquiry, working out what it is about and passing it to the right person or list, so the ones that matter are easier to spot. In a busy week, a good enquiry sitting unread in the wrong inbox is one of the easiest ways to lose a customer.

Drafting replies for a person to check. A first reply written from your own information, prices and policies. Someone on your team reads it, changes what needs changing and sends it. The value is not the draft itself. It is that the reply starts from something instead of a blank screen.

Organising customer information. Pulling the useful details out of emails and forms and into one tidy record, instead of someone typing them in again.

Summaries and follow-up. Summarising a long thread before a call, and flagging the enquiry that has gone quiet, so follow-up no longer depends on somebody remembering.

First drafts of routine content. Social posts, product descriptions, the email that goes out every month. We do this ourselves: Cheers, the social media tool we built, suggests drafts with AI, and a person checks every post before it is published. It runs at The Anchor, our own venue.

All five have one thing in common: none of them changes what you sell. They change how much of your week goes on the admin around selling it.

Where it is a distraction

AI because AI is in the brief. If the starting point is "we should be using AI", the project usually ends with a tool nobody uses. Start with the job.

Automating a process that does not work yet. If enquiries already go missing, AI will lose them faster. Fix the route first, then speed it up.

Taking the person out where judgement matters. Complaints, refunds, anything with a legal or safety side, and the conversation that decides whether a customer comes back. These are where a person earns their keep.

A chatbot in place of a clear website. If customers keep asking the same questions, the first fix is usually a page that answers them plainly. A chatbot answering questions your website should have answered is a sign the website needs work.

The question underneath the question

"Should we be using AI?" is almost never the question that is actually holding a small business back. Underneath it there is usually one of these:

• Not enough of the right people find you.
• People enquire, but too few become customers.
• The prices or the offer are wrong for the customers you want.
• The owner is doing work that someone, or something, else should be doing.

Only the last one responds directly to AI. The first three are about demand, conversion and what you sell. AI can make the work on them faster, but it cannot decide it for you, and speeding up the wrong thing just gets you to the wrong place sooner.

Three questions before you buy any AI tool

1. Which job takes the most hours each week, and does it have a clear right answer? If you cannot name one, you are not ready to buy anything. If you can, that is your first use.
2. Who checks the output, and what happens when it is wrong? Every AI tool gets things wrong some of the time. Decide in advance which results a person approves before they reach a customer, and which are safe to run on their own.
3. What information would it need to read, and are you allowed to give it that? Customer details, prices, policies, past emails. Check where the tool processes data and whether it learns from what you give it. The Information Commissioner's Office guidance on AI and data protection is the place to start. When we checked on 5 November 2026, the ICO said that guidance was under review following the Data (Use and Access) Act, so look for any update.

If all three have clear answers, try AI on that one job, with one person responsible for how it goes, for a month. Then decide whether to keep it, change it or drop it, before you add a second use.

If you would rather work through it as a set of statements than as three questions, the AI readiness assessment asks twelve of them about how your business actually runs, and gives an honest read on where AI would help and where it would not.

What we will and will not claim

We do not publish figures for how many hours AI gives back, because we cannot prove a number that would hold for your business.

What we can point to is our own venue, The Anchor, where the results came from changing commercial decisions rather than adding technology. Rebuilding the website around what people search for grew Google Search visibility by 828%. Changing how bookings were confirmed and reminded cut no-shows by 89%. AI made the research and analysis behind those decisions quicker. It was not the reason the numbers moved.

The Anchor is a pub, but the problems it had are the ones every small business has: being found, turning interest into a firm booking or sale, and keeping more of what you sell. AI makes a good decision quicker to act on. It does not make the decision for you.

For one sector worked through in detail, read AI for accountants.

The short version

AI belongs in a small business, in specific places, at the repetitive end, with a person checking the result. That is worth doing, and it is not particularly hard.

But it is rarely the thing standing between a business and its next stage of growth. Work out what is, first. If it turns out a repetitive job really is eating your week, our AI integration work starts by agreeing the task, the data and the checks before anything is built. And if something simpler would do the job, we will say so.