Public sector organisations across the UK remain wary of AI’s ability to deliver near-term productivity improvements, despite sustained investment and political pressure to use the technology to drive efficiency.
That’s according to new research from Snowflake, based on a survey of 500 senior decision-makers conducted by YouGov.
The findings highlight a widening gap between ambition and execution, with only 23 percent of UK organisations reporting productivity gains from AI at scale, and 45 percent seeing benefits limited to pilots or narrow use cases.
The report lands as ministers continue to position AI as central to economic growth and public service reform, including ambitions to boost the UK economy by £47 billion annually through adoption.
Public sector takes cautious approach
The research suggests the public sector is taking a more cautious, governance-led approach to AI, with longer timelines expected for productivity gains than in some other sectors.
More than half (52 percent) of public sector leaders said AI will not materially improve productivity for at least two years, while 66 percent reported that ethics and safety considerations significantly shape adoption decisions.
Concerns around reliability are also prominent, with 53 percent citing safety of AI outputs as a key barrier to confidence.
This caution, while aligned with regulatory and public accountability requirements, may delay the realisation of efficiency gains compared to less regulated sectors.
Structural barriers outweigh technology challenges
Across all sectors, the report finds that the main obstacles to scaling AI are organisational rather than technical.
If you liked this content…
Skills shortages, poor data quality, siloed teams and unclear ownership are cited as the primary barriers to progress. Only 19 percent of respondents identified the technology itself as a major constraint.
Responsibility for AI is often fragmented across senior leadership, IT and data teams, limiting accountability and slowing decision-making – a challenge that is particularly acute in complex public sector organisations.
Just 24 percent of respondents said their AI initiatives are prioritised using a clear framework aligned to organisational objectives.
Investment continues despite limited returns
Despite these challenges, confidence in AI remains high. Most organisations expect to increase investment over the next 12 to 24 months, with just one percent planning to reduce spending.
Cost reduction is currently the dominant measure of success, cited by 44 percent of respondents, compared with 26 percent who prioritise revenue growth.
Dr Fabian Stephany, who contributed to the research, said the findings reflect a typical lag between technological breakthroughs and measurable productivity gains.
He noted that organisations need time to adapt workflows, governance models and workforce skills before benefits are realised at scale.
“AI systems are only as powerful as the people who develop, apply and govern them,” he said, pointing to ongoing shortages in AI-related skills as a critical constraint.
