SolarWinds' survey found that UK public sector IT teams are rolling out AI across fragmented systems, highlighting a gap between adoption plans and technical readiness.
The global survey of 1,040 IT professionals found that 57% of public sector respondents work in hybrid environments, while 25% operate primarily on premises. According to SolarWinds, that mix creates a difficult foundation for organisations trying to introduce AI tools across services.
Fragmentation emerged as a central issue, with nine in 10 respondents reporting fragmented IT tooling across systems and platforms. That can complicate efforts to deploy AI consistently.
The data also suggests the operational strain is already visible. Some 26% of respondents said integrating AI into existing systems is a key source of increased complexity, while 69% said they need to double-check AI outputs and 67% said they struggle to trust its recommendations.
Keeping pace with the technology is another pressure point. Nearly a third, or 31%, said keeping up with AI tools and capabilities is a major challenge.
At the same time, some teams said they are being pushed to move faster than their organisations can support. Around 16% said they feel pressure to adopt AI faster than their organisation is ready for, while 42% said better infrastructure would help them adapt to AI demands more quickly.
Mixed estate
The findings add to the wider debate over how public sector bodies should manage AI adoption in ageing, complex technology estates. Hybrid environments, on-premises systems and fragmented toolsets can make it harder to standardise data, oversee outputs and connect newer applications with existing workflows.
For public sector organisations, those constraints can carry extra weight because services often depend on older systems that are difficult to replace quickly. The survey suggests AI is exposing those weaknesses rather than removing them.
Rich Giblin, Head of Public Sector and Defence at SolarWinds, said: "AI is creating real opportunities for public sector organisations to work more efficiently, improve services and make better use of their data. But only if the right foundations are in place first. The real risk is moving faster than your data, systems and processes can support.
"If AI is being introduced into environments where information is siloed, systems don't connect cleanly, or knowledge is outdated, then the technology will reflect those weaknesses rather than fix them. In practice, that can mean unreliable outputs, poor recommendations and much more effort spent checking and validating results."
"That's why the safest and most effective approach is to start with well-defined use cases in areas where organisations have stronger data, clearer processes and better oversight. From there, AI can be expanded more confidently. But if public sector organisations try to scale adoption before the groundwork is done, they risk limiting the value AI can deliver across services."
The results show that concerns are not limited to infrastructure. Trust in AI output remains a practical obstacle, with more than two-thirds of respondents saying they either need to verify results or struggle to rely on system-generated recommendations.
That matters because the public sector has been under sustained pressure to improve productivity and service delivery while managing tight budgets and legacy estates. In that context, AI tools may offer efficiency gains, but the survey suggests many IT teams believe the underlying environment still needs work before those gains can be realised at scale.
The figures on hybrid and on-premises systems also indicate that fully modernised cloud environments remain far from universal across the public sector. As a result, IT teams may be asked to fit new AI applications into estates that were not designed for them.
SolarWinds' data suggests better infrastructure remains one of the clearest requests from public sector technology teams. Among respondents, 42% said improved infrastructure would help them adapt to AI demands faster.