UK firms spend on AI without clear returns, study finds

Air IT Group has published research on what it describes as an AI decision gap among UK businesses. The study found that 83% of decision-makers had either spent on AI without measurable returns or were not measuring returns at all.

Based on 504 responses from decision-makers at UK organisations, the research examined why companies are investing in AI, where projects are stalling, and whether businesses can show value from that spending. In financial services, the pattern was more pronounced: 89% reported at least some AI spending without measurable returns, or said returns were not measured, while 43% said their organisation had introduced an AI tool before agreeing the business problem it was meant to solve.

Across the full sample, 41% said their organisation had bought or introduced an AI tool before agreeing what problem it should solve. Among respondents with at least one active AI initiative, 74% said not all initiatives had both an agreed business objective and a measure of success.

The findings point to a gap between AI adoption and formal planning. While many organisations are under pressure to act on AI, the data suggests a significant share are doing so before setting clear goals for what they want systems to achieve or how results will be assessed.

The issue also appears in the number of projects that have been dropped. Among respondents whose organisations formally track AI pilots, 62% said at least one pilot had been abandoned in the previous 12 months.

Integration problems and staff uptake were among the main reasons. Some 34% cited integration difficulties as the most common reason for abandonment, while 29% pointed to employee non-adoption.

Measurement problems

The research also suggests that, even where AI is in use, measuring its effect on work remains difficult. Almost half of respondents, 47%, said AI had made a task or process take longer once checking, correction, setup, and management were taken into account.

Only 13% said their organisations consistently convert time saved through AI into measurable increases in output or productivity. That suggests time savings alone do not necessarily translate into business gains unless companies change workflows, staffing, or output expectations in response.

Financial services respondents reported a similar pattern of weak measurement and uneven returns. More than one in five, 22%, said more than half of their AI spending over the previous year had delivered no measurable return.

The figures come as businesses face growing pressure to show practical outcomes from digital investment. AI spending has risen across many sectors, but evidence of returns has often been harder to produce, particularly where projects begin as experiments or are deployed without clear ownership within the business.

Air IT said the findings showed the importance of defining the business problem before introducing a tool. It argued that organisations should also agree in advance how success will be measured, so leaders can decide whether to extend, revise, or stop a project.

"AI has huge potential for businesses, but simply introducing more tools doesn't guarantee better outcomes. Our research found that 41% of organisations had introduced an AI tool before agreeing what problem it should solve. If the problem and the outcome aren't clear from the outset, it becomes much harder to know whether AI is actually delivering value," Lee Johnson, Chief Technology Officer at Air IT Group, said.

Johnson added: "Businesses need to start with what they're trying to achieve, identify where AI can make the biggest difference and agree how success will be measured. That gives leaders the evidence to decide what to scale, what to rethink and where to invest next."

Wider backdrop

The study reflects a broader tension in the UK market between enthusiasm for AI and the practicalities of implementation. Businesses may feel competitive pressure to adopt new tools quickly, but the survey indicates that governance, integration, and workforce adoption continue to shape whether those tools produce measurable results.

For sectors such as financial services, where compliance, legacy systems, and process controls can complicate technology roll-outs, those challenges may be more visible. The sector's higher share of respondents reporting unmeasured or weak returns suggests AI spending is not immune to the same scrutiny as other technology budgets.

Air IT is headquartered in Nottingham and provides IT, cyber security, cloud, data, automation, and AI services to UK organisations. It supports more than 2,000 organisations from nine regional offices with round-the-clock operational coverage.