UK firms see weak AI returns as skills lag adoption

Businesses are seeing limited returns from artificial intelligence investments because workforce skills are lagging behind adoption, according to QA. The training company cited research showing that formal AI training remains uneven across organisations.

The study found that 32% of employees had received no formal AI training, while only 15% said they had ongoing or advanced support. Advanced use also remained limited, with about 9% of workers describing themselves as advanced or expert users.

Most employees were operating at a basic level. Around 24% said they used AI only for simple tasks such as drafting emails, summarising documents, producing meeting notes or rephrasing text.

That pattern appears to be limiting returns on spending. UK businesses are investing an average of £235,000 per company in AI and emerging technology, yet only 16% of employees reported significant productivity gains.

A further one in 10 employees said they could achieve more with AI but lacked the training or support to do so. The findings suggest a gap between deploying AI tools and enabling staff to use them in ways that have a broader effect on operations.

Uneven adoption

AI use also varied sharply by role. Technical staff in IT led in advanced usage, while employees in administration, operations, customer service and sales were more likely to use AI only for basic tasks or not at all.

QA attributed that divide partly to lower confidence and limited access to training. As a result, gains in productivity and efficiency are concentrated in a small group rather than spread across the workforce.

Dr Vicky Crockett, Portfolio Director for AI at QA, outlined what organisations need to do before expecting broader returns from AI programmes. “Before diving into a full AI transformation, organisations need to build basic AI and data literacy so everyone feels confident using these tools. It’s also essential to provide role-specific training, because AI affects different jobs in different ways and a one-size-fits-all approach simply doesn’t work.

“By tailoring upskilling to individual roles, businesses can maximise the value AI brings to everyday tasks. Finally, developing internal AI champions helps create momentum, as early adopters can share insights and support colleagues as they adapt to new ways of working,” Crockett said.

Skills focus

QA’s findings add to a wider debate over whether corporate AI spending is moving ahead faster than workforce preparation. Businesses across sectors have introduced generative AI tools into daily work, but many are still working out how to train staff beyond introductory use.

QA’s research highlighted a clear distinction between access and impact. AI tools may now be common in organisations, but for a sizeable share of employees, use remains concentrated on low-impact activities.

That matters for companies seeking measurable returns on investment, particularly when spending has already reached substantial levels. If only a small minority of workers can use AI in more sophisticated ways, the business case for broader deployment may remain difficult to prove.

Jo Bishenden, Chief Learning Officer at QA, said companies should treat AI as a workforce issue, not just a technology rollout. “AI is being adopted at pace, but too many organisations are still treating it as a technology rollout rather than a shift in people capability.”

“There’s a growing gap between what AI is capable of and how it’s actually being used at work. As AI evolves towards more agentic models, value no longer comes from basic use or high-level guidance, but from equipping people with the skills, confidence and judgement to work effectively alongside these systems.”

“The organisations seeing the greatest productivity gains are those investing in capability building at scale, embedding AI skills into everyday roles and enabling their people to apply AI in ways that genuinely improve how work gets done,” Bishenden said.