Public sector facilities teams unprepared for AI, Bellrock survey finds

TL;DR:

• A Bellrock survey of 285 estates professionals across UK local government and NHS estates found that just 1% believe their current facilities management model is fit for purpose, 40% lack the skills to deploy AI, and 65% have received no AI training at all.
• Confidence in property data quality is also low: 3% of local government professionals and 6% of healthcare professionals are very confident their data is accurate, complete and standardised enough for AI to be effective.
• Resultsense view: this complements yesterday’s SolarWinds survey of UK public-sector IT teams and confirms a wider readiness gap. AI strategies in Whitehall and the NHS will sit on top of estates and infrastructure teams that lack both the data and the training to absorb them.

The survey, published by property services firm Bellrock, focuses on the operational layer of UK public-sector AI adoption rather than the policy or vendor-platform layer that dominates most coverage. Estates teams manage the physical environments — schools, hospitals, council buildings, depots — that any wider AI strategy will need to instrument and operate.

What the readiness gap looks like

Legacy systems are explicitly named as a brake: 75% of local government professionals and 80% of healthcare professionals say outdated technology is affecting AI progress moderately, significantly or severely. Only 7% of local government workers and 3% of healthcare workers fully agree that their organisation has the in-house skills to deploy AI effectively in estate management. Senior leaders are unconfident too — only 3% say they are very confident using data and AI insights for strategic property decisions.

Cost and investment constraints were cited as the top barrier to AI adoption, followed by skills and capability gaps, lack of quality data, legacy systems and infrastructure, leadership culture, and risk and governance concerns.

Where the appetite is

There is broad support for a standard data taxonomy across each sector — 68% of local government professionals and 80% of healthcare professionals agree or strongly agree that one is needed. That is a relatively easy intervention compared with the workforce skills gap, and it is consistent with the Bank of England, ICO and Cabinet Office push for shared metadata standards across regulated sectors.

UK context

The findings sit alongside the SolarWinds survey of UK public-sector IT teams published last week — 56% of which said AI had made their roles more demanding rather than easier — and the public-sector AI training scheme rollouts announced by the Government Skills and Curriculum Unit earlier this year. Together, the picture is one of expanding AI ambition meeting compressed budgets and unprepared frontline teams.

Looking forward

Without targeted training, data taxonomy work and a path off legacy estate-management systems, AI strategies that depend on the estates layer — energy, occupancy, building safety, predictive maintenance — will struggle to convert pilot wins into operational savings. UK local authorities and NHS trusts that move first to address the data quality and training questions will be the ones positioned to capture AI’s productivity dividend in their physical estates.