SME owners across London are fielding AI questions from clients and increasingly from their own staff. AI for SMEs has moved from a future consideration to a live decision. UK productivity growth barely moved in the year to the first quarter of 2026, up just 0.4% according to the ONS, and AI is one of the few tools being pitched as a way to lift it. For most SMEs, the question is now how to introduce AI without wasting money or creating a data problem.
The Risks of AI Adoption Without a Plan
The Department for Science, Innovation and Technology surveyed 3,500 UK private-sector businesses with five or more staff for its AI Adoption Research. The two most commonly cited barriers to adoption were a lack of identified use for AI (71%) and limited AI skills in the business (60%). Cost sits further down the list at 23%, though for businesses that do face it, 76% rate it as significant.
Few of these barriers are really about the technology. They’re about planning. The same research also had a limitation. It could only capture AI use that business leadership already knew about. Tools staff are already using independently don’t show up in these figures at all. Once businesses were using AI, data security and the accuracy of its output became their biggest concerns. For a business handling client data or anything covered by a confidentiality agreement, the real risk sits between what staff are approved to use and what they’re doing day to day.
What AI for SMEs Looks Like When It’s Done Properly
Most businesses that had adopted AI spent around a year considering it before switching anything on, according to the same DSIT research. The Lloyds Bank Business Barometer, published in March 2026, found that London businesses using AI report the strongest results in the UK: 93% saw a rise in productivity and 71% reported higher profits. The most common spending bracket was under £25,000, which a third of adopters stayed within.
That kind of result tends to start with solving a specific problem, such as slow admin or an overloaded inbox. The problem gets picked before the tool does. It also comes from someone checking the output. DSIT found that 84% of businesses using AI apply at least some human oversight to what it produces. For an SME, that means being clear on the problem you’re solving, on who checks the output, and on where your data ends up.
Three Things to Put in Place Before You Spend Money on AI
Know where your data goes
Before any tool touches client information or anything covered by a confidentiality agreement, check what the provider does with that data and whether it trains their model on it. The ICO’s guidance on AI and data protection sets out what UK businesses are expected to have in place. Checking a new tool against it takes five minutes.
Put something in writing about who can use what and for what
Very few small businesses currently have any internal policy on AI use. Staff end up making their own calls on what’s safe to paste into a tool. A short, plain-English policy sorts this out without slowing anyone down. Our AI and automation page covers the questions we’d ask before switching anything on, including who has access to your data and whether staff are already using tools outside company policy.
Set a number before you start, even a small one
The most common spending bracket in the Lloyds research was under £25,000, and those businesses still reported a return. A modest budget is enough to test a tool properly. Decide on a figure and a deadline, then assign someone to judge whether it paid for itself.
Talk It Through at Summer Sessions: AI, Unlocked
We’ve supported London SMEs with their IT for over two decades, and AI is now becoming part of those conversations too.
If you’d like to explore what sensible AI adoption could look like for your business, join us at Summer Sessions: AI, Unlocked on Thursday 29 October at the South Place Hotel.
Dan and the team will run a short, practical session on how SMEs can approach AI with a clearer plan, including where it can genuinely help, what to think about before introducing new tools, and how to avoid unnecessary risk.
There’ll also be drinks afterwards and time to talk things through with other business owners facing many of the same questions.
Register your interest here and bring along any questions you have about using AI in your business.
FAQs
Is AI worth it for a small business right now?
For most SMEs, the answer depends on whether there’s a specific task worth automating. Where there is, the return usually follows. Without one, adoption tends to produce nothing worth measuring.
How much does it cost a small business to start using AI properly?
There’s no fixed figure. Most UK businesses currently using AI report spending relatively modest sums to get meaningful results, often through existing subscriptions like Microsoft 365 Copilot. Starting small and measuring the result is the part that counts.
Will AI reduce headcount at my business?
Most UK businesses already using AI don’t expect it to change the size of their team. DSIT’s research found it’s used most often to remove repetitive tasks and reshape how work gets done.
