Councils are already stretched: more residents need social care, more children need an Education, Health and Care Plan, and complaints and information requests keep landing faster than teams can clear them. That pressure is real, but it is also where the opportunity sits. The same AI that is adding to some of that volume can help a council process it too, and the capacity that frees up is what actually changes a council’s trajectory.

Why do councils always feel like they’re reacting?

Councils feel like they are permanently reacting because demand is rising on several fronts at once. More residents need social care. More children need an Education, Health and Care Plan. More complaints, requests and enquiries land through every channel a resident can use, and now one extra driver is making that queue grow faster still. Researchers at the Centre for the Governance of AI and the Hertie School call it agentic flooding: AI agents make it cheap enough to produce a detailed request that the volume and complexity landing on a public service climbs sharply. Their study logged 84 cases across 11 jurisdictions, and UK examples are easy to find.

The Housing Ombudsman has seen complaints climb from around 1,700 a year to roughly 7,000, as tenants use chatbots to draft longer, more detailed cases. The Bar Standards Board has recorded a rise of about 25% in complaints against barristers, with AI-generated reports taking longer to assess. Employment tribunals now carry a backlog of more than 64,000 open single cases, with AI-drafted claims adding to both the volume and the complexity. Agentic flooding is only one driver among several, but it is a clear sign that the reacting side of a council’s workload keeps growing, not shrinking.

What is the difference between managing demand and preventing it?

Demand management and demand prevention solve different problems, and mixing them up is where a council AI project can lose its way. Demand management means handling what already lands more efficiently: triaging it, drafting a first response, absorbing a higher volume without adding headcount. Demand prevention means reading across every channel a resident uses, calls, complaints, web forms, and fixing the underlying issue before the request is ever made.

Prevention is the bigger prize. A council that can see a spike in complaints about a service and fix it before the formal requests follow saves far more than it ever will by getting faster at answering them. But prevention needs a joined-up view of contact across the whole organisation, and most councils do not have one yet. Building it takes time, data work and a service team that can spare the attention, which is exactly what is in short supply when a service is already underwater.

The honest sequence is to relieve the pressure first. Free up officer time on what is already arriving, and that capacity becomes the on-ramp to the bigger, more transformative work.

What does relieving pressure actually look like?

Relieving pressure looks like handing officer time back on work that is already arriving, and a Freedom of Information request is a clear example of how that plays out. Take FOI as one example, not the only one: the two jobs inside it are well understood, triage a request when it arrives, then analyse it and draft a response, and both are now candidates for AI support that keeps a person in charge of every decision.

UCL research puts the average FOI request at 7.5 hours of officer time, around £187.50 at the statutory rate and closer to £250 to £300 once real staff cost is counted, against a 20 working day deadline. That is the time an officer gets back for every request handled this way.

On the resident-facing side, AI can help someone search what a council has already disclosed, check whether a similar request has already been answered, and submit a new one directly, cutting down on repeat and already-answered questions before they become casework. On the officer-facing side, AI can triage an incoming request, work through a 100-page submission, and draft a response, with any exemption surfaced for a person to review rather than applied automatically. Every response is signed off by an officer before it is sent, and that decision is tied to a named officer in the audit trail. The case system a council already runs stays the record of what was decided and why, and the AI runs inside the Microsoft 365 environment the council already has, so there is no new system to migrate to and no new supplier holding the data.

The same shape of support is already showing up elsewhere. In children’s services, AI-assisted drafting is handing caseworkers back 75% of the time they used to spend drafting an Education, Health and Care Plan, on average. FOI is simply the clearest example to walk through end to end, because the triage, drafting and sign-off steps are easy to follow from request to response.

How does freed-up capacity lead to bigger change?

Freed-up officer time is the capacity a council needs to start on demand prevention: pulling complaints, calls and web contact into one view, spotting the pattern behind a spike in requests, and fixing the cause rather than processing the next thousand symptoms. That is a longer piece of work, and it becomes realistic once a team has bought itself some room, rather than staying permanently behind on what is already arriving.

We have seen this play out at one council currently working through rapid prototyping across several service areas, including its contact centre and complaints handling, with the aim of building exactly this kind of joined-up view. It is early, and the data foundation that view depends on is still being built rather than delivered, but the direction is the right one: relieve the pressure that is easiest to see, then use the time it buys to go after the harder, more valuable problem.

“AI and specialist tech has reduced administrative burden and improved consistency.”

— Councillor Keith Cunliffe, Deputy Leader, Wigan Council

We build your capability, not your dependency, so a council keeps that view and keeps improving it, with or without us involved.

Key takeaways

  • Council demand is rising across several fronts at once, and AI-generated volume, what researchers call agentic flooding, is one driver among them: Housing Ombudsman complaints, Bar Standards Board reports and employment tribunal claims have all risen sharply.
  • Demand management (handle what arrives faster) and demand prevention (stop it arriving) are different problems, and sequencing them, management first, matters.
  • FOI is one clear example: it costs a council roughly 7.5 officer hours per request, and AI support built inside the council’s own systems hands most of that back without moving any data anywhere new.
  • The same pattern already shows up elsewhere, including EHCP drafting, where caseworkers get back 75% of drafting time on average, and every draft response in this pattern is signed off by a named officer, nothing sent automatically.
  • Freed-up officer time is the on-ramp to demand prevention, not the end goal in itself.

Frequently asked questions

What is the difference between demand management and demand prevention?

Demand management handles requests that have already arrived more efficiently, for example triaging and drafting FOI responses faster. Demand prevention looks across every contact channel to catch and fix the underlying issue before a resident ever needs to submit a request. Management buys the time and evidence a council needs before it can realistically tackle prevention.

Is AI making council demand worse, or better?

Both, depending on where it sits. AI is one driver behind rising complaints and requests, what researchers call agentic flooding, because it makes a detailed submission cheap to produce. The same technology, used on the council’s side, can triage and draft responses, handing officer time back rather than adding to the backlog.

Does creating capacity mean cutting jobs?

No. Freeing up officer time is not about reducing headcount, it is about redirecting the hours already lost to repetitive triage and drafting toward harder work: reading across services, spotting patterns, and preventing demand before it arrives. Every decision, including any FOI exemption, still sits with a named officer.

Sources