Public sector · Social housing
70% of UK public sector organisations say their data landscape is fragmented. Inside a housing association, that fragmentation now has regulatory teeth.
Those numbers from the 2026 UK public sector data trends report describe most housing associations we work with. They explain why Awaab's Law evidence is harder to assemble than it should be, why TSM submissions still need manual rework, and why boards see plenty of compliance activity but struggle to see strategic data progress.
Key takeaways
70%
Say data is fragmented, not interoperable
58%
Cite skills shortages as top barrier
£39bn
SAHP grant against milestone reporting
£45bn
Public sector modernisation savings unrealised
Housing associations don't operate alone. Tenants move between associations and local authority allocation lists. Vulnerable households are visible to multiple services before any one team raises a concern. Repairs go through external contractors. Safeguarding requires data flowing between social care, schools, police and health.
The pillar report names safeguarding, hospital discharge, and housing-and-prevention as the cross-agency use cases where governed data sharing pays back fastest. Snowflake's secure data sharing model removes the physical movement of data between organisations, which is the biggest blocker most information governance leads cite.
Housing associations are explicitly named in the 2026 trends report as a growth area for cloud modernisation. The reasons are practical: legacy estates, scarce engineering capacity, and a board appetite for visible efficiency without large transformation programmes.
Cloud-native platforms offer consumption-based economics and faster time to value. For an association that cannot staff a large engineering team, that matters more than the architecture diagram.
This is a real architectural choice. Most associations are deeply embedded in Microsoft, with Power BI and increasingly Fabric on the table. The instinct is to default to the Microsoft estate because licences are already in place.
| Housing workload | Microsoft Fabric | Snowflake |
|---|---|---|
| Data sharing with LAs, ICBs, contractors | Limited | Strong fit |
| Predictive damp and mould risk models | Constrained by capacity units | Elastic compute, strong fit |
| Cost predictability under sector pressure | Capacity-based, can creep | Consumption-based, transparent |
| Small in-house team ownership | Fine if Microsoft-first | Fine either way |
Power BI as a visualisation layer is fine. The question is what sits underneath it.
For an association where every pound on data infrastructure is a pound not spent on housing people, cost transparency matters more than the architecture diagram.
Housing data is sensitive: tenancy, vulnerability, safeguarding, financial hardship. Governance is no longer a compliance exercise. It is the strategic capability that decides whether your association can share data safely with local authorities and partners, evidence Awaab's Law decisions, and use AI without creating new risk.
Only 26% of the public believe government uses AI responsibly. That trust gap is real, and it shows up the moment a tenant asks how their data is being used.
Only 26% of governments globally have deployed analytics or AI partially or fully. The reason is rarely the AI itself. It is the underlying data: incomplete, inconsistent, fragmented, undocumented.
For housing associations, the most valuable AI use cases are predictive: damp and mould risk by property profile, rent arrears segmentation between cannot-pay and will-not-pay, EPC retrofit prioritisation against asset, cost and health impact. None of those models work on data that is not connected first.
Treat compliance as the entry point. Awaab's Law and TSM data is the use case that justifies investment now.
Connect before you visualise. A new dashboard on broken data does not pass a Regulator's audit.
Frame everything in tenant outcomes. Housing associations are charities. Every pound on data infrastructure is a pound not spent on housing people.
Build it as a knowledge transfer engagement. Your team has to own it long after the partner leaves.
Prove it before you commit
A 30-day Snowflake Proof of Value on your priority workload (an Awaab's Law evidence trail, a TSM dataset, or a repairs feed). Two days of Catalyst BI consultancy. Real cost numbers. No multi-year commitment.
Book a POV scoping callKeep reading
For the full picture across the UK public sector, including all eight trends and the supporting research:
2026 UK Public Sector Data Trends →