UK firms adopt AI widely but shallowly, ONS finds

TL;DR:

• Self-reported AI use among UK businesses with 10 or more employees has risen from around 12% to around 35% since late 2023, but the average number of AI technologies per adopting firm has moved only from 1.4 to 1.6.
• Just 10% of adopting businesses describe their use as extensive, and only 15% say more than half their employees use AI daily.
• The ONS reads this as pointing to “limited transformative impacts for most AI-adopting firms”.

Three years of adoption data now exist, and the shape they describe is a mile wide and an inch deep. The Office for National Statistics has published its first combined view of AI in UK businesses covering 2023 to 2026, and the headline adoption figure — a near-tripling — sits alongside a depth measure that has barely moved.

Large language models lead at 18% of businesses, followed by visual content creation at 16%, machine learning data processing at 12% and image processing at 6%. Robotics reaches 2%.

Efficiency, not new products

Close to 60% of businesses report using AI to improve existing operations. Fewer than one in five use it to develop new products or services or to reach new markets — the applications with more room to compound. Boston Consulting Group’s Centre for Growth director Raoul Ruparel told the Financial Times the efficiency focus was expected but “far from the most likely approach to generate long-term returns”, while noting that microbusinesses were more inclined towards revenue-generating cases.

That inversion matters for smaller UK firms. Businesses with fewer than 10 employees report higher extensive use (17%) than those with 250 or more (9%), and are likelier to apply AI to new products and markets. Scale is not translating into depth.

The binding constraint is expertise

Among businesses reporting barriers, lack of expertise is the most cited, peaking near 18% for firms with 100 to 249 employees. Cost affects 7% to 14%. Notably, 41% report no barriers at all — the constraint is less regulatory friction than knowing what to do. Firms citing an expertise gap are far likelier to train existing staff (62%) than those reporting none (26%), yet only 11% of businesses say more than half their workforce has had AI training. That gap echoes what London’s largest employers reported this week.

Employees are ahead of employers: 55% report using AI for work or education against 35% of businesses reporting a deployed technology — much of it informal, unmanaged and invisible to the firm.

Looking forward

Employment effects remain muted, with around half of adopters reporting no headcount change, consistent with Bank of England Decision Maker Panel findings from January. The exceptions are specific: over half of firms using visual content AI report effects on creative and design roles. For UK businesses hoping AI will lift stubborn productivity growth, the ONS numbers suggest most have bought a tool, not changed how they work.