UK businesses are continuing to pour money into AI despite most failing to achieve meaningful AI productivity gains in the UK at scale, according to new research released today by Snowflake.
The study, conducted by YouGov on behalf of Snowflake and surveying 500 senior UK business decision-makers, found that just 23% of organisations have achieved AI-driven productivity improvements at scale, while a further 45% say gains remain limited to specific or experimental use cases.
Despite this, appetite for AI spend shows no sign of cooling. The research found only 1% of organisations plan to reduce AI investment over the next 12 to 24 months, suggesting confidence in AI’s long-term potential remains firmly intact even where short-term results have proved elusive.
The findings land at a moment of significant policy focus on AI as an economic lever. The UK Government’s AI Opportunities Action Plan aims to boost the economy by £47 billion annually, estimating that widespread AI adoption could increase national productivity by up to 1.5% each year. For organisations yet to act, the window may be narrowing.
For more on why UK organisations are struggling to turn AI investment into measurable outcomes, read our analysis of the key AI and automation trends shaping 2026.
Internal barriers, not technology, are slowing AI productivity UK-wide
The report challenges a common assumption that technology readiness is what holds organisations back. Only 19% of respondents cited technology as a barrier to progress. Instead, the primary obstacles are skills shortages, poor data quality, organisational silos and unclear strategic direction.
Governance also emerged as a structural weak point. Just 24% of organisations say AI initiatives are prioritised using a rigorous framework aligned to business objectives, meaning the majority of deployments lack clear strategic grounding. Responsibility for AI governance is typically fragmented across executive, technology, data and business leaders, with no single clear owner: a recipe for slow decision-making and limited accountability.
This pattern is consistent with broader research into AI’s impact on UK workplaces, which has found that organisations investing in AI without strong governance frameworks often see effort intensify rather than reduce.
Dr Fabian Stephany, Economist and Departmental Research Lecturer at the Oxford Internet Institute, University of Oxford, said the findings were consistent with historical patterns around transformative technology. He commented:
“Technological breakthroughs rarely translate immediately into productivity improvements, as organisations need time to adapt their workflows, governance structures and capabilities.”
Dr Stephany also pointed to skills as a critical and growing constraint, drawing on his SkillScale research group’s findings that workers with AI-related skills already command a wage premium of around 23% in the UK, alongside better job prospects and additional benefits. For organisations that delay building those capabilities, the talent gap (and the productivity gap) is only likely to widen:
“Expanding access to AI skills and training will be critical if organisations want to sustain and scale these productivity gains.”
Which UK industries are winning and losing the AI productivity race
The research highlights notable differences in AI maturity across UK industries. Financial services organisations are more advanced on governance and strategic alignment, though regulatory and reputational concerns are slowing the move to scale. Manufacturing firms express strong belief in AI’s long-term potential but anticipate slower returns due to skills gaps and integration challenges. Retail lags on both confidence and delivery, with AI frequently confined to isolated use cases amid persistent data quality issues and fragmented ownership.
The public sector presents perhaps the most cautious picture. Some 52% of public sector leaders say AI will not materially improve productivity for at least two years, with 66% reporting that ethics and safety concerns significantly shape adoption decisions, and 53% citing the reliability of AI outputs as their top concern. While that caution reflects a responsible approach to risk, it also risks leaving significant efficiency gains unrealised as other sectors move faster: a tension already visible in wider workplace analytics data for 2026.
Cost cutting over growth: How UK businesses are measuring AI success
When it comes to measuring AI’s value, cost reduction leads the way. Nearly half of respondents (44%) cited it as the most important measure of success, ahead of revenue growth at 26%. Around 40% of all organisations surveyed expect AI to take two years or more to deliver material productivity improvements.
These findings chime with UC Today’s own analysis of how UK organisations are evaluating AI platforms in 2026, which found that buyers are increasingly demanding proof of operational gains rather than accepting vendor promises at face value.
Jennifer Belissent, Principal Data Strategist at Snowflake, said the research pointed to a clear gap between ambition and execution. She said:
“Productivity gains require clear ownership, strong data foundations and alignment between AI initiatives and measurable business objectives. The focus must now shift from experimentation to disciplined execution.”
For UK organisations still finding their footing with AI, the message from Snowflake’s research is pointed: the technology is ready. The question is whether they are.
