Ask the owner of a small UK business what they would do with an extra day a week and the answers are rarely about technology. They want more time with customers, more time to quote for new work and fewer evenings spent on invoices, emails and spreadsheets. That last group is where artificial intelligence is starting to make a practical difference, well away from the headlines about robots and chatbots.
The numbers show how quickly things have moved. According to the Office for National Statistics, the share of UK businesses with ten or more employees using at least one AI technology rose from around 12% in late 2023 to around 35% by June 2026. In the same analysis, 28% of the smallest firms, those with up to nine employees, reported using at least one AI technology.
Yet the same ONS analysis describes that adoption as relatively shallow, which suggests many firms are experimenting rather than changing how they work. The rest of this piece looks at why, where AI tends to pay off first in a small business, what UK data protection law expects once customer information is involved, and when outside help is worth paying for.
What the ONS figures say about how small firms use AI
The headline growth hides a quieter detail. Between late 2023 and mid-2026, the average number of AI technologies used by each adopting business rose only modestly, from about 1.4 to about 1.6. In other words, more firms have tried AI, but the typical adopter still uses only one or two AI technologies. Large language models, the technology behind tools such as ChatGPT, were the most widely used type.
Asked what holds them back, businesses in earlier ONS analysis most often cited difficulty identifying business use cases, cost and a lack of expertise. That matches what many owners say privately: they have opened an AI chatbot, asked it a few questions, and then struggled to see how it connects to the work that actually fills their week.
The admin jobs AI handles best
The most reliable early wins sit in the back office. Supplier invoices can be read and their details extracted for bookkeeping software. Incoming emails and web enquiries can be sorted and answered in draft. Meeting notes can be turned into follow-up tasks. Quotes can be assembled from previous jobs. None of it is exciting work, which is why it so often waits until late at night.
The tasks worth automating first share three traits. They happen often, they follow a recognisable pattern and a mistake is easy to spot before it causes harm. A plumbing firm sorting fifty enquiries a week, a lettings agent chasing paperwork or an accountancy practice processing receipts all fit the description. A once-a-year task, however painful, rarely justifies the setup time.
Before buying anything, spend a week noting how long these jobs take and who does them. That baseline is what lets you judge, a month later, whether the automation has saved time or merely moved the work somewhere else. It also tends to reveal that the real problem is messy data, such as duplicate customer records, which needs fixing first.
What UK data protection law expects
Most of the admin that AI handles well involves personal data: customer names, addresses, phone numbers and sometimes much more. The UK GDPR applies however small the business. One of its core principles, data minimisation, requires personal data to be adequate, relevant and limited to what is necessary for the purposes for which it is processed.
For AI, that principle translates into a practical question for every task: what does the tool actually need to see? To sort enquiries by type, it does not need a customer’s date of birth. To summarise a complaint, it does not need their bank details. Taking names and reference numbers out of a document before it goes into a general chatbot is often the simplest way to stay on the right side of the principle without losing the benefit.
The Information Commissioner’s Office publishes detailed guidance on AI and data protection, and small firms will find its explanations useful. The ICO’s own page notes that this guidance is under review following the Data (Use and Access) Act 2025, so it is worth checking the current version on the ICO website before relying on any single point.
Your staff are already using AI
In many small businesses, AI arrived before anyone decided to adopt it. A team member uses a free chatbot to rewrite a customer email, summarise a contract or tidy up a spreadsheet, and pastes in whatever the task requires. That is understandable, and often productive, but it means customer information may be leaving the business through tools nobody has checked.
A short written policy helps more than a ban. Name the tools that are approved, say which kinds of information must never go into anything else and explain how to strip out personal details when a general tool is the only option. Keep it to a page, discuss it at a team meeting and revisit it when the tools change, which they will.
When to bring in outside help
Simple automations are within reach of most owners, often using features already built into accounting, email or booking software. Projects that link several systems, read documents reliably or draft messages to customers are harder, and that is usually where specialist help becomes worthwhile. The ONS finding about a lack of expertise is a fair reflection of the problem.
Help does not have to be local, although distance makes the data questions matter more. Aivy, a team of AI automation specialists in Australia, works with the tools a business already uses and begins with a free 30-minute discovery call. Whoever you speak to, ask where your customer data will be processed, who maintains the system and how its results will be measured.
Small steps, measured carefully
Small firms that make AI stick tend to follow the same pattern. They choose one repetitive task, measure it before and after, keep personal data to the minimum the task needs and only then move on to the next one. That approach takes a little longer to start, and it is more likely to still be running a year later.
For a small UK business, the real prize is time. Every hour not spent keying in invoices or answering the same question for the tenth time is an hour available for customers, staff and the work that brings in revenue. Starting with the admin is the most dependable way to win that time back.
