AI tools are now part of everyday work in a growing number of UK businesses, mostly for admin, marketing and writing. The ones getting real value start small: one repeat task, clear rules on what information can go in, and a person who checks what comes out.
AI use doubled in a year
The government’s latest AI adoption survey found that one third of UK businesses were using at least one AI technology in 2026, up from 16% in 2025. Fewer businesses now think AI is irrelevant to them, too: 38%, down from 51% the year before (Department for Science, Innovation and Technology).
The Office for National Statistics tells a similar story. Among businesses with 10 or more employees, AI use rose from around 12% in late 2023 to around 35% by June 2026. But only 10% of those using it said they use it extensively (Office for National Statistics).
So more businesses are trying AI, but most are still dipping a toe in. That matches what I see. Plenty of owners have used ChatGPT or Copilot to write an email. Far fewer have built it into how the business runs.
Why it keeps changing
Part of what makes AI hard to keep up with is that it does not arrive as one product. The assistants themselves change every few months, and AI features keep appearing inside software you already pay for, such as your email, documents, accounts package or workflow system.
That is good news, because you may not need anything new. It also means a tool your team tried and dismissed a year ago may now do the job, and a tool someone signed up for on their own may be holding client information nobody knows about.
What businesses are using it for
Among businesses using or planning to use AI, the government survey found administration was the most common use (81%), followed by marketing (71%) and IT (66%). Tools based on large language models, the technology behind assistants such as ChatGPT, Gemini and Copilot, were by far the most common (Department for Science, Innovation and Technology).
In a small business, the useful jobs tend to be:
- drafting emails, quotes and proposals from a few notes
- summarising meetings, long email threads and documents
- turning a messy description of a task into a first draft procedure
- sorting incoming enquiries and pulling the key details into a system
- writing first drafts of website pages and social posts for a person to edit
None of these remove the need for someone who knows the business. They remove the blank page and the retyping.
Where it goes wrong
The same survey shows the common problems. Limited AI skills affected 67% of businesses already using it, and 56% had concerns about accuracy and bias (Department for Science, Innovation and Technology).
In practice, I see three ways it goes wrong. People trust an answer that sounds confident but is wrong. Client or staff information gets pasted into a free tool without anyone checking what happens to it. And AI gets pointed at a process that was never agreed, so it produces inconsistent work faster.
Personal data needs particular care. UK GDPR applies to personal information whether a person or an AI tool is handling it, and the ICO has published guidance on what that means in practice (Information Commissioner’s Office).
A sensible way to start
- Pick one task that happens every week, follows roughly the same steps and currently eats time. Writing up meeting notes or first replies to enquiries are good candidates.
- Write down how the task should be done before you involve AI. If the team cannot agree the steps, fix that first.
- Decide which tools are allowed and what information can and cannot go into them. One short page is enough.
- Try it for two or three weeks with a named person checking every output.
- Compare the time taken before and after, including the checking. Then keep it, change it or drop it.
Once one task works, the next is easier, because the team has a pattern to follow and the rules are already written.
Getting help with it
You do not need an AI strategy to get value from these tools. You need a clear process, the right tool for that process and a team that knows how to use it safely. That is the work we do: we look at where the time goes in your business, pick the tasks where AI and automation will make a real difference, set the tools up around how your team works and train people to use them. If you would like a second opinion on where to start, get in touch.
