AI for small business: what's useful today and what's just hype?

Hardly a week goes by without a new claim that artificial intelligence (AI) will transform your business or put you out of it. If you run a joiner's firm in Paisley, a dental practice in Glasgow or a charity shop in Ayr, it is hard to know what deserves your time. This guide separates the practical uses from the noise, and explains the limits and privacy checks that matter.

What people actually mean by AI

In everyday business use, AI usually means software that has learned patterns from huge amounts of existing material and can then produce something similar or spot something familiar. The tools most owners meet are chatbots that write text when you type a request. This is called generative AI, because it generates new wording, images or summaries rather than just retrieving what already exists.

It is not a thinking colleague. It predicts what a sensible answer looks like. That is why it can be very helpful and also confidently wrong, sometimes within the same reply.

Where it is useful today

The best uses share a pattern: the task is routine, a person can check the result quickly, and a mistake would be caught before it does harm. Some examples that fit:

• Drafting a first version of a routine email, such as a polite reminder to a customer who has not paid an invoice, which you then edit in your own voice.
• Summarising a long document, such as a lease, a supplier contract or a set of meeting notes, so you know which parts to read closely.
• Turning rough notes into a tidy job description, a menu description or a newsletter outline.
• Transcribing a voicemail or recorded call so it can be searched and filed.
• Suggesting wording for a website page or social media post that you then check for accuracy and tone.
• Sorting incoming enquiries into categories, such as quote requests, complaints and general questions, so the right person sees them first.

Notice that in each case a person stays in charge. The tool saves the blank-page problem or the repetitive sorting, not the responsibility for what goes out under your name.

Where the hype outruns reality

Be wary of anything promising to run a part of your business on its own with no oversight. Claims that a tool will guarantee more sales, replace your whole admin team or work perfectly from day one are marketing, not evidence.

There are some real limitations to keep in mind:

• Invented facts. AI tools sometimes state things that are untrue, including made-up prices, regulations or references. The industry calls this hallucination.
• No knowledge of your business. Unless you supply the details, the tool does not know your opening hours, your service area or your terms.
• Out-of-date information. A tool may not know about recent changes in law, tax or local rules.
• Generic wording. Text produced without careful guidance can sound like everyone else's, which does little for a local firm that trades on personality and trust.

Privacy and data protection

This is the area where small businesses most often get caught out. If you type customer names, health details, addresses or financial information into a public AI tool, you may be sharing personal data with a third party. Under UK data protection law, you remain responsible for how that information is handled. The Information Commissioner's Office (ICO), the UK regulator, publishes guidance on AI and data protection.

Healthcare practices, solicitors, accountants and charities handling vulnerable people's details need to be especially careful, because the information is sensitive and clients expect confidentiality. If a customer-facing tool talks to the public, say so plainly so nobody thinks they are chatting with a member of staff.

Checks to make before you rely on it

1. Read the output properly. Check names, figures, dates and promises before anything is sent or published.
2. Check the supplier's terms. Find out where your data goes, how long it is kept and whether you can switch off the use of your inputs for training.
3. Decide who is accountable. Name one person responsible for each AI-assisted task.
4. Keep a record. Note what you used a tool for, so you can explain it to a customer, insurer or regulator if asked.
5. Review after a month. Ask whether it saved time, whether errors crept in, and whether customers noticed a difference.

Starting small and sensibly

Pick one dull, repeatable task and try a tool on it for a month. A café owner might draft the weekly specials email. A plumber might summarise customer messages at the end of the day. You will learn far more from that than from a grand plan.

Some practical help is also available when you want something more permanent. A website chatbot, for example, can answer common questions such as opening hours or service areas outside working hours, provided it is set up with accurate information and a clear route to a real person. And a CRM (customer relationship management system, which simply keeps customer details and conversations in one place) gives you tidy, well-organised information, which makes any later automation far safer to use.

If you would like a second opinion on where technology could genuinely help your business, On The Spot offers a free digital strategy review, with no obligation to act on it.

In short

AI is useful today for first drafts, summaries, transcription and sorting routine work, as long as a person checks the result and customer data is protected. It is not a substitute for judgement, it can be confidently wrong, and the bigger promises are mostly hype.

Next steps

1. Choose one routine task that takes up time each week and write down how it is done now.
2. Try one AI tool on that task for a month, using no personal or confidential information.
3. Read the supplier's data terms and the ICO's guidance on AI and data protection.
4. Check every output before it reaches a customer, and note any errors you find.
5. Review the results with your team and keep, change or drop the tool.