Quick answer: For narrow, repetitive tasks — drafting, summarising, admin — probably yes, and often at no cost: in the UK government's 2026 survey, 47% of businesses that could estimate their AI spending reported spending nothing, because free tools or features in software they already pay for did the job. But keep expectations honest. UK adoption doubled in a year (16% to 32%), yet only half of adopters reported an overall productivity rise, those gains were described as modest, and 64% saw no immediate change in operating costs. The firms that get value use AI on named tasks, keep a person checking the outputs, and measure hours saved on one process at a time. Try one task for 30 days before spending anything.
What the 2026 numbers actually say
Three official or research sources published this year give a fairly consistent picture of the UK.
The Department for Science, Innovation and Technology (DSIT) published its AI adoption survey report 2026 on 29 October 2026, based on telephone interviews with 3,503 businesses between January and April. In 2026, 32% of UK businesses were using at least one AI technology — double the 16% recorded in 2025. Adoption was uneven: 52% among larger firms and 61% in information and communication, against a 39% average in business services.
The Office for National Statistics (ONS) paints the same direction of travel. Its analysis of AI in UK businesses (released July 2026) puts self-reported AI use at around 35% of businesses with 10 or more employees in June 2026, up from around 12% in late 2023.
And the British Chambers of Commerce, working with the University of Essex (report, March 2026), found 54% of surveyed firms actively using AI — a sample that was around 94% small and medium-sized businesses — up from 35% in their 2025 survey.
| Source (published) | Headline figure | Detail worth noting |
|---|---|---|
| DSIT survey (Oct 2026) | 32% of UK businesses use AI, double 2025's 16% | 3,503 businesses interviewed Jan–Apr 2026 |
| ONS analysis (Jul 2026) | ~35% of businesses with 10+ employees, up from ~12% in late 2023 | Average AI technologies per adopting business rose only from ~1.4 to ~1.6 |
| BCC & University of Essex (Mar 2026) | 54% of surveyed firms (mostly SMEs) actively use AI | 95% of SME users report no change in workforce size; 86% report unchanged job roles |
If you are still working out where AI might fit in your own operation, the free practical guide on the site is a calmer starting point than any statistic: read the guide.
The part most headlines skip
The same DSIT report contains numbers that matter more than the doubling headline — especially if you run a small business:
- Gains are real but modest. Half of adopters reported an overall productivity increase, and the report describes those gains as "generally described as modest". The most common benefits were task automation and tidier workflows, not transformation.
- Costs barely moved. 24% of adopters reported reductions in operating costs; 64% saw no immediate financial effect; 9% saw costs increase.
- Use is shallow. The ONS found the average number of AI technologies per adopting business has barely moved since 2023 (roughly 1.4 to 1.6). Most firms use one or two things, not a stack.
- Only 14% of businesses consider AI of high or critical importance to their operations — despite a third of them using it.
- It is cheap to join. 47% of businesses that could estimate their AI spending reported no expenditure in 2025. Free tiers and AI features inside existing software carry most of the current use.
- Where it lands: administration (81% of businesses using or planning to use AI), marketing (71%), IT (66%) — accessible, low-complexity work. Natural-language tools, mostly: 77% of those using or planning AI involve large language model approaches.
- The barrier is not price. Limited skills, expertise or understanding of use cases affected 67% of adopters and 78% of businesses planning to adopt. Among non-adopters, 65% cited a lack of clear business case or proven return.
- People stay in the loop. 81% of businesses apply some human oversight to AI outputs — a sensible habit, even if the report notes it slipped slightly since 2025.
Read together, the honest summary is this: UK businesses are adopting AI widely, using it lightly, getting small real gains, and spending little to do it. Nobody's business model has been redefined — including among the adopters.
So is it worth it? How to judge for your own business
National averages cannot answer that question; a named task in your own week can. It is usually worth trying AI on a task when all four of these hold:
- The task repeats weekly or more (drafting replies, summarising documents, first-pass admin).
- The inputs are already digital and can be anonymised if they are sensitive.
- You or a colleague can check the output in a couple of minutes.
- The process itself works — AI on a broken process just produces mistakes faster. If the underlying workflow needs rework rather than speed, that is usually an automation and process question to settle first.
And it is usually not worth bothering when: nothing in your week repeats enough to matter; outputs cannot be checked by someone who knows the work; or the data involved is too sensitive to share even in anonymised form.
If those conditions hold, run a simple 30-day test:
- Name one task. Not "AI for the business" — one task, e.g. first drafts of supplier email replies.
- Baseline it. Note how long the task took last week, roughly.
- Start free. Use a free tier or the AI feature in software you already pay for.
- Keep a human check. Every output gets read before it goes anywhere.
- Compare after 30 days. Hours spent, quality of results, how it felt.
- Decide. Hours clearly down and quality acceptable — keep it, maybe add a second task. No real change — stop; you have spent curiosity, not money.
A fictional illustration, labelled as such: imagine a two-person bookkeeping practice where chasing client documents eats three hours a week. A 30-day test of AI-drafted reminder emails — checked by a human before sending — might cut that to two hours. That is a modest, real gain of the kind the national data describes, and it cost nothing to discover. If the test had shown no difference, the practice would have stopped with the same information for free.
If you would rather judge this against your own business than against national averages, the Business Process & AI Diagnostic page explains what the fixed-fee review covers: up to three agreed, bounded processes, plain-language written findings and a prioritised list of improvements — including cases where a simple process change beats AI — at £349 as a one-off (€399 in Portugal). If you want to build the checking habit first, a Personal AI Session works through one task you bring, in English or Portuguese. And if you would simply like to talk it through, the free 30-minute fit conversation is optional and commits you to nothing.
The wider picture: the EU and Portugal
The same pattern — fast growth from a low base, modest reported gains — holds across Europe. Eurostat reports that 20.0% of EU enterprises with 10 or more employees used AI technologies in 2025, up from 13.5% in 2024. In Portugal, INE's survey of ICT use in enterprises found 11.5% of companies with 10 or more people in service used AI (2025 survey) — below the EU average, which mostly says the opportunity is still open rather than exhausted.
One Portugal-specific note: the PRR funding line "IA nas PME", which co-financed AI adoption projects for SMEs, has closed — its official programme page is currently marked as closed ("Fechado"), with no announced reopening. The practical consequence is that an AI decision should stand on its business case — hours saved on a named task — rather than on subsidy arithmetic.
Frequently asked questions
Is AI worth it for a small business in 2026?
For specific, repeated, checkable tasks — yes, often, and usually starting free. As a general business transformation, the 2026 evidence does not support big claims: half of UK adopters report modest productivity gains, and most see no immediate cost change.
Do I have to pay for AI tools to get value?
Not to start. 47% of businesses that could estimate their AI spending reported spending nothing, and most current use runs on free tools or features inside software businesses already pay for.
Will AI reduce headcount in a small business?
So far, the evidence says no: 95% of UK SMEs using AI reported no impact on workforce size over the past year, and 86% reported unchanged job roles (BCC/University of Essex, March 2026). The 2026 pattern is AI as a support tool, with people responsible for checking its work.
Last updated: 22 November 2026