Why 95% of UK Councils Are Exploring AI — And What's Stopping the Other 5%

The headline is striking: 95% of UK councils are using or exploring AI. That figure, from the Local Government Association's State of the Sector Survey in 2025, reflects a sector that has moved decisively from curiosity to commitment. But the same survey reveals a harder truth: only 36% have realised productivity benefits. The gap between exploration and impact is the central challenge for local government in 2026.

This article looks at what the data actually shows — and what councils that are breaking through are doing differently.

The State of Play

The LGA's numbers paint a clear picture of adoption. 83% of councils are already using generative AI in some form. 68% are using external AI tools (from chatbots to analytics). That's no longer early adopter territory; it's mainstream. Councils are experimenting with AI for customer contact, back-office automation, and policy analysis.

Yet 36% benefits realised means that for most, the link between investment and outcome is still weak. Pilots run, tools are trialled, but the step change in productivity or resident experience has not yet materialised. The reasons are not lack of ambition. They are structural: funding, capability, and capacity.

The Three Barriers

Funding (62%) — Budget pressure is the most cited barrier. AI projects often require upfront cost: licences, integration, training. In a climate where every pound is scrutinised, finding a budget line for "AI" is difficult. The business case has to be crystal clear and tied to savings or quality that finance and members can support.

Capability (56%) — Many councils report that they lack the in-house skills to evaluate, procure, and manage AI systems. Digital teams are stretched; data and AI literacy is uneven. Without someone who can ask the right questions of vendors and hold them to account, projects drift or underdeliver.

Capacity (52%) — Even when funding and capability exist, day-to-day capacity is thin. Running a pilot, managing change, and iterating on the basis of feedback all take time. Officers are already overloaded; adding "lead the AI project" to the list often means it slips.

These three barriers reinforce each other. No funding means no capacity to explore. No capability means business cases don't get written. No capacity means even good tools don't get embedded. Breaking the cycle requires a different approach.

How Leading Councils Are Breaking Through

Councils that are in the 36% — or clearly on the path — tend to share a few characteristics.

Start with a clear problem. AI works best when it's solving a defined problem: contact centre deflection, service navigation, case triage. The London Borough of Newham, for example, deployed GovAI to help residents find the right support across financial, housing, and employment services. The problem was clear; the success metric (fewer avoidable contacts, better outcomes) was measurable.

Choose platforms that don't demand a big IT project. The more a solution requires custom integration, migration, or rip-and-replace, the more it runs into the capacity and capability barriers. Embeddable tools that work on existing websites and systems reduce risk and time to value.

Invest in data sovereignty from day one. Councils that have navigated procurement and compliance successfully have often chosen UK-hosted, transparent AI. When the DPO and procurement can see where data sits and how it's used, barriers come down.

Partner for capability. You don't need to build everything in-house. Partnering with a provider that understands local government — and can support implementation, change, and iteration — turns capability from a barrier into a lever.

The 95% figure will only grow. The question is how many councils will move from exploring to realising benefits. The ones that do will be those that pick a clear problem, choose tools that fit their capacity, and partner for the capability they need. The data is clear; the path is there.