Half of UK SMEs now use AI, and most are only using it for content

UK SMEs have crossed the halfway line on AI. According to the British Chambers of Commerce's March 2026 Future of Work research, 54% of firms are now actively using AI, up from 35% in 2025 and 25% in 2024. That is the fastest jump the BCC has recorded. But the same research shows most of that usage is shallow: generic chatbots and writing assistants drafting content, summarising documents and pulling together research. The interesting gap in 2026 is no longer between firms that use AI and firms that don't. It is between the half using the same tools in the same way, and the roughly one in ten who have built AI into how the business actually runs.

"AI adoption stopped being an edge the moment it became the average. Half the market is drafting the same passable blog post with the same tool. The one in ten who wired it into how the business runs are the ones pulling away."

Dean Cookson, founder, Operosus

What does the data actually say?

The BCC study, run with Atos and analysed with the University of Essex's Centre for Micro-Social Change, surveyed firms of which around 94% were SMEs. Three findings stand out.

First, adoption has gone mainstream. At 54%, using AI is now the norm, not the exception. If your competitors are UK SMEs, the majority of them have some form of AI in the building.

Second, depth is rare. The full report finds that most firms are using generic tools to support tasks like drafting content, summarising information, analysing data or conducting research. Only around one in ten report adopting bespoke AI systems designed for their organisation, the kind integrated into workflow automation, forecasting, operational decisions or customer service.

Third, the believers are the users. SMEs already using AI report strong net expectations of productivity improvement (+71 percentage points in the BCC's measure), while firms still on the fence are far less optimistic. Experience, not theory, is what converts people.

Earlier polling points the same way on what AI is used for. YouGov's survey of 1,000 UK SME decision-makers found that among adopters, 54% use AI for automating tasks and 45% for marketing and advertising, with operational areas like customer service (31%) and logistics (28%) trailing behind. Content and marketing remain the front door through which most firms enter. (Every UK adoption figure in this piece is kept sourced in our UK small business AI statistics table.)

Why does everyone start with content?

Because it is the easiest win. Writing a product description, a LinkedIn post or a first-draft proposal in a chat window requires no integration, no data work and no process change. You open a tab, you type, you paste the result. The barrier is close to zero, which is exactly why half the market has already walked through it.

That is also why it confers almost no advantage. When your competitor can produce the same passable blog post with the same tool in the same afternoon, faster mediocre content is not a moat. It is the new baseline.

None of this means content tools are a waste of time. They save real hours, and the BCC data suggests they do so without disruption: 95% of generic AI users report no impact on headcount, because these tools help existing staff move faster rather than restructure anything. The point is simply that "we use ChatGPT" is now a description of the market, not a strategy.

Where has the advantage moved?

To the one in ten. The BCC report draws a sharp line between generic users and firms running bespoke AI systems built around their own operations. Bespoke adopters are significantly more likely to report changes in how the business is structured and staffed, in both directions, because the technology is actually load-bearing. It sits inside the quoting process, the customer pipeline, the scheduling, the follow-up, not alongside them in a browser tab.

That distinction maps to something we see constantly at Operosus. The difference between a tool and a system is whether the work flows through it. A writing assistant produces an output someone still has to do something with. A system takes the enquiry, qualifies it, drafts the response, routes it to the right person, follows up and records what happened, with people supervising the judgement calls rather than pushing every button.

Our own products exist because of that gap. Bidwell does not just write tender text; it manages the work of finding, qualifying and answering UK tenders end to end for small firms. Contentwell and Emailwell sit inside a pipeline that takes content from draft through review to publication and follow-up, rather than leaving a pile of AI drafts in a folder. And the bespoke systems we build for clients connect lead capture, CRM, email and reporting into one flow, because that is where the hours actually disappear: not in writing the first draft, but in everything around it.

What should an SME do differently?

Start from the bottleneck, not the tool. The firms getting compounding value from AI did not begin by asking "what can we do with ChatGPT?". They asked where the business loses the most time or leaks the most revenue, then worked out whether AI could carry part of that load reliably.

Three questions sort this out quickly:

1. Which process eats the most skilled hours on repetitive work? Tender responses, quote follow-ups, inbox triage and reporting are common answers. These are system candidates, not chat-window candidates.
2. Where does work currently fall through the cracks between tools? If a lead arrives by web form, gets retyped into a spreadsheet and chased by memory, the win is connecting those steps, with AI handling the drafting and routing inside the connection. Our guide to moving from spreadsheet to system is the diagnostic for exactly this.
3. What would you check before trusting the output? Whatever your answer is, that check belongs inside the system as a review step. The BCC findings on training point the same way: firms that invest in structured AI training and deeper systems are the ones reorganising around the technology rather than bolting it on.

The honest summary of the 2026 data is encouraging for any SME owner who feels behind: you are probably not. Half the market is at roughly the same stage, using roughly the same tools for roughly the same tasks. The firms pulling away are not the ones with the cleverest prompts. They are the ones who picked a process that matters, wired AI into it properly and let the gains compound while everyone else was drafting another blog post.

If you want a second opinion on which of your processes would repay that treatment first, that is exactly the conversation we have with SMEs every week as part of our custom build work. Get in touch and we will tell you, plainly, whether there is a system worth building. And for what each new model release actually changes for this picture, see what the latest models change for SMBs.