This time last year, using AI in UK local authorities was largely about cautious experimentation. I am sure most of you reading this who tinkered with AI in its early days found the tools basic, clunky, and limited in what they could do.

Over the past year, however, the situation has changed. AI is now a practical and flexible solution that can be used across many services. Believe it or not, several UK councils have shown clear value from AI in areas such as social care, planning, highways, governance, and CRM casework. At the same time, techies and governance personnel have moved quickly to strengthen ethics, transparency and equality safeguards to ensure that innovation and responsible use can progress together.

 

Quick summary – how has AI grown in 2025?

2024 was the year that large language models made big improvements in reasoning, writing code and understanding language. These breakthroughs laid the groundwork for AI to move into real, everyday organisational use in 2025.

Throughout 2025, one clear message emerged from industry analysis: AI stopped being a novelty and became a practical work tool. McKinsey reported that AI agents can now handle tasks such as processing payments, checking for fraud, and identifying issues automatically without constant human involvement. IBM also highlighted this shift, noting that more capable and autonomous AI systems moved beyond proof-of-concept trials and into real, live deployments.

 

How has AI impacted local authority implementations in 2025?

A survey by the Local Government Association, carried out between December 2024 and February 2025, found that around nine in ten UK councils are already using AI or actively exploring it. Councils are running both public-facing pilots and internal tools across services such as highways, social care, planning and enforcement. At the same time, analysis by the Tony Blair Institute shows the scale of the opportunity. Its modelling found that AI could improve or automate around 26% of council tasks, saving around one million staff hours and roughly £30 million per year in a single authority. Scaled across England and Wales, this could represent up to £8 billion in annual savings.

 

Of the many implementations of AI in 2025, here are some standout examples

St Helens Borough Council created an in-house Technology Enabled Care (TEC) Hub using Microsoft Copilot Studio to help match residents with the most suitable care technology, automatically drafting personalised care plans to support discharge and reablement. The team identified clear, cashable savings from the use of automated medication dispensers (around £160,000 saved for every 20 residents) and reinvested these savings to expand the council’s AI and TEC capability. The next phase, planned for 2026, is to connect this work to a resident-facing digital front door through the council’s chatbot, Helen, making TEC easier for residents to discover and access.

Medway Council introduced Lilli, an AI-driven in-home monitoring system that uses sensors to spot changes in daily routines and to flag potential risks early. The system has helped reduce care costs and delivered around £1.6 million in savings within its first year. Building on this success, the council has extended the use of Lilli to support people with learning disabilities, helping more residents live safely and independently at home.

Barking and Dagenham Council developed an in-house AI tool called BD Notes to help draft statutory case documents. Using the tool, the time needed to draft sections of Education, Health and Care Plans (EHCPs) has fallen from four to five hours to under five minutes, while officers remain fully responsible for reviewing and approving the final content. The council plans to embed BD Notes across 13 service areas by 2026 and expand the range of templates it supports, allowing the tool to be rolled out quickly to new areas of work.

Of course, AI adoption has not been without challenges. In some cases, high-value contracts with suppliers such as Palantir have attracted public and political scrutiny, highlighting the importance of transparency and robust data protection assessments when introducing AI into council services.

 

With 2026 fast approaching, what is the next phase of AI development for councils?

AI will move from being simple “helpers” to acting more like digital workers. These digital workers will be able to handle routine tasks on their own (e.g. checking rules, drafting responses, updating systems), and only passing things to a human when approval or judgement is needed. This is especially valuable for councils, as they deal with large volumes of repeat processes every day.

Organisations will increasingly use lots of small, specialised AIs, each focused on doing one job well, rather than relying on one large AI system to do everything. These smaller AIs can work together as a system. This approach is easier to control and govern, easier to change or switch off if something goes wrong, and it costs less than running one large, all-purpose AI.

AI will also become much better at understanding more than just text. It will be able to look at photos of potholes or fly-tipping, listen to phone calls and turn them into summaries, and read scanned forms or handwritten notes. Because councils deal with many different types of evidence, this will help them process work more quickly and more consistently.

Councils and their suppliers will increasingly adopt smaller, more focused models, rather than using huge global AI models. These will be trained around UK law, local policy and service guidance, making them more accurate, safer to use and easier to manage.

AI will continue to be built directly into everyday software, instead of feeling like a separate tool. We are already seeing this with Microsoft Copilot inside emails and documents. For council staff, this means they will not need to “learn AI” as it will simply become part of how they already work. At the same time, stronger rules, checks and safety controls will become standard. This includes clear human oversight, data protection safeguards and checks to make sure AI is used fairly and responsibly.

Support from central government will also increase, with active programmes to help councils adopt AI in a safe and consistent way, rather than each authority working in isolation.

 

Conclusion

As we move into 2026, the focus is shifting toward AI systems that work together, models designed specifically for council services, faster modernisation of older systems, and clear, auditable controls to make sure AI is used safely. Together, this marks a change in how AI is viewed: no longer as a one-off experiment or pilot, but as shared public-sector infrastructure that supports how UK councils deliver services every day. The councils that act now to modernise their approach will be those best equipped to meet the challenges of the coming decade.