The debate about whether councils should use AI is largely over. In 2025 the Local Government Association found that 95% of councils were already using or exploring it, though only 7% considered themselves leaders using it at scale. The more useful question now is where AI is actually earning its place, and what councils get back when it does. Here are five areas where it is earning its place in 2026, ordered roughly from most established to most emerging.

1. Drafting Education, Health and Care Plans

One of the most established examples sits in children’s services, where AI drafts Education, Health and Care Plans for SEND teams. It is a strong case because the work is statutory, high-volume and under real pressure: the Department for Education recorded 718,800 active EHCPs in January 2026, up 12.5% in a year.

AI that drafts EHCPs is now live in 40 councils and returns around 80% of the time caseworkers used to spend writing. North Somerset brought it in to handle a 95% rise in EHCP demand. The signal that matters most is not the percentage but the caseworker who said they could not go back to working without it.

2. Speeding up adult social care assessments

Adult social care is the second area where AI has reached genuine day-to-day use, mostly in drafting Needs Assessments and case notes. These are long, repetitive documents that pull social workers away from the people they are meant to support, which is exactly the kind of task AI handles well.

In Wigan, more than 300 social workers now use AI daily, and drafting time on Needs Assessments has dropped by around half. Time saved on writing is time returned to visits and relationships, which is where the value actually lands.

3. Taking pressure off the contact centre

The contact centre is where most residents meet the council, and it is under constant strain. In Wigan, the contact centre handles more than 1,000 calls a day across 22 service lines, with advisors working across as many as 11 screens.

That scale, and the time advisors lose moving between systems, is why it is one of the areas Wigan has been prototyping AI: summarising a caller’s history, routing the query, and drafting responses so advisors spend less time hunting and more time resolving.

It is a high-potential use precisely because it compounds. A small saving on each of a thousand daily interactions adds up quickly, and residents get a faster answer.

4. Clearing planning and regulatory backlogs

Planning is a quieter application but a telling one, because it is procedural and error-prone. In Wigan, the service handles around 1,500 planning applications a year, and roughly 40% fail validation the first time. AI can check an application against requirements before it enters the queue, catching the gaps that would otherwise bounce back and cost weeks.

The same pattern holds across regulatory casework, from building control to licensing. Anywhere a case has to be checked against a fixed set of rules, AI can do the first pass and leave officers to focus on judgement rather than validation.

5. Automating back-office and information work

The most common use of AI in councils is the least visible. The LGA found the corporate and back-office layer, covering HR, administration, procurement, finance and cyber security, is where 84% of councils using or exploring AI have applied it first, ahead of adult social care at 44% and children’s services at 31%.

This is automated meeting minutes, procurement and finance admin, the work that keeps a council running but rarely gets attention. In Wigan, automated meeting minutes alone saved 107 days of work.

Freedom of information requests are a good example of the pattern. They are time-bound, repetitive and rising, and AI can draft responses and pull the right material far faster than a manual search. It is unglamorous, and it is where a lot of the early time savings are found.

Where this leaves councils

The through-line across all five is that AI works best where the task is high-volume, rules-heavy or document-heavy, and where the saving frees people for the parts of the job that need a human. Two of these, EHCP drafting and adult social care assessments, are already running in live services across multiple councils. The others are being built now, in programmes like Wigan’s, against problems councils have measured and understand.

To be straight about who is behind this piece: Agilisys builds AI inside real councils, with the frontline staff who use it, and the standard we hold ourselves to is that the system keeps working with or without us. We build your capability, not your dependency, the data stays in your tenancy, and your team can run what we leave behind. The point is not to add another tool but to give people back the time these five examples show is recoverable.

If you are weighing where to start, pick the workflow that is both high-volume and painful, and prove it in one service before scaling. The councils seeing results did not try to do everything at once. They started where the time saving was obvious and built from there.

Key takeaways

  • Councils are using AI most in back-office work, adult social care and children’s services, according to the LGA’s 2025 research.
  • AI works best where the task is high-volume, rules-heavy or document-heavy: EHCPs, social care assessments, contact centre queries, planning checks and FOI responses.
  • The ones live across multiple councils are the safest starting point: EHCP drafting in 40 councils returning around 80% of drafting time, and adult social care assessments with 300+ social workers in Wigan using AI daily.
  • Start with one high-volume, painful workflow and prove it in a single service before scaling.

Frequently asked questions

What are councils actually using AI for in 2026?

Mostly high-volume, document-heavy work. The most common areas are back-office and corporate admin, adult social care and children’s services, according to the LGA’s 2025 research. Concrete examples include drafting EHCPs, adult social care assessments, contact centre responses, planning validation and freedom of information requests.

Is AI in councils actually saving time?

Yes, where it is deployed on the right tasks. AI that drafts EHCPs returns around 80% of caseworker drafting time across 40 councils, and in Wigan automated meeting minutes saved 107 days of work while Needs Assessment drafting time dropped by about half. The savings come from repetitive writing and checking, not from replacing judgement.

Where should a council start with AI?

Start with a workflow that is both high-volume and painful, and prove it in one service before scaling. Statutory, document-heavy areas like SEND or adult social care tend to show a clear saving fastest, because the work is repetitive and the pressure is already high.

Sources