Published 27/08/2026
Author: Chris Aslett
AI is making it easier for citizens to generate longer complaints, requests and enquiries. Copilot Agents could help councils respond without adding more pressure to already-stretched teams.
The answer is not to use AI to decide which citizens should be heard. It is to use AI to help council teams understand, filter, prioritise and respond to the growing volume of information coming through the door.
Recent reporting has highlighted a growing challenge for public bodies: citizens are using generative AI to create more detailed and sophisticated correspondence. Used well, that can be positive. AI can help residents articulate concerns, understand their rights and navigate services they may previously have found difficult to access.
But when straightforward issues arrive wrapped in pages of explanation, references and follow-up questions, the administrative burden increases. Officers still need to read the request, work out what is being asked, find relevant information, check existing cases, identify escalation points and respond.
The problem is not AI-generated complaints
It is important not to frame this as a battle between councils and citizens. The challenge is the pace and complexity of demand. Councils are being asked to process more information, more quickly, with the same limited capacity.
This is where Microsoft Copilot Studio and Copilot Agents become particularly interesting for local government. Rather than thinking about AI simply as a chatbot, councils can start thinking about agents as digital workers that carry out defined processes using approved knowledge, clear instructions, governance rules, connected systems and escalation points where a human must take over.
At CPS, we have explored this opportunity through an FOI Agent designed to support councils managing Freedom of Information requests. When an FOI request arrives, an agent could help with the initial processing before an officer begins detailed review.
It could read the request, identify what information is being asked for, break complex wording into specific requirements, flag ambiguity, search approved sources, identify information that may already be public, route the request to the right team and prepare a response pack for the FOI officer.
The point is not for an agent to decide whether information should be disclosed. That remains a human decision. The value is in removing the administrative weight around that decision so officers can focus on judgement, compliance and accountability.
Now imagine this across the council
The FOI Agent is just one example. The same model could support many council processes where teams face high volumes of correspondence, fragmented information and limited time.
could identify the core issue, categorise the complaint, detect whether it relates to an existing case, summarise the history, highlight urgency or safeguarding concerns, identify relevant policies and prepare a case summary for the complaints officer.
could group representations around common themes, identify duplicate or similar submissions, extract the substantive planning issues and help officers focus on material considerations. The goal would not be to decide whether an objection should influence a planning decision, but to help planners see the substance of large volumes of feedback more clearly.
could help residents get to the right answer or service more quickly. For example, it could explain why a council tax bill has changed using approved information, or recognise that a missed bin enquiry relates to an existing service request and route it accordingly.
This is where Copilot Agents become more than an efficiency tool. If councils can process incoming information intelligently, they can start to spot patterns that are hard to see when everything sits across thousands of emails, forms and cases.
An agent might show that 600 enquiries relate to the same service issue, 250 complaints point to one recurring process failure, 180 FOI requests ask for information already published, or hundreds of planning representations cluster around just a few themes.
That changes the conversation from “How can we answer more enquiries?” to “What are these enquiries telling us about our services?”
500 planning representations → 5 recurring themes
The art of the possible
The most interesting thing about Copilot Agents isn’t necessarily what councils can automate today.
It is what becomes possible when you start designing AI around specific public-sector processes.
Think of a future council with a network of specialist agents:
| Agent | What it could help with |
|---|---|
| FOI Agent | Understand, classify and prepare FOI requests |
| Complaints Agent | Triage, summarise and route complaints |
| Planning Agent | Analyse and group planning representations |
| Customer Services Agent | Answer routine enquiries and guide residents |
| Housing Agent | Understand repairs and housing enquiries |
| Revenues Agent | Help interpret council tax and benefits enquiries |
| Member Enquiries Agent | Prepare briefings and identify relevant case history |
| Consultation Agent | Analyse large volumes of public responses |
| Case Management Agent | Summarise histories and prepare cases for officers |
These don’t have to operate independently.
They can become part of an agentic operating model, where specialist agents handle defined parts of a process and hand work to humans when judgement, accountability or expertise is required.
This isn’t about replacing council officers
And this is perhaps the most important point.
The objective isn’t to create a council where citizens talk to machines and officers disappear behind them.
It is almost the opposite.
Let the agent read the 20-page submission.
Let the agent identify the three issues that actually matter.
Let the agent find the relevant policy.
Let the agent identify the existing case.
Let the agent prepare the summary.
Then let the human make the decision. That is where AI can create genuine capacity.
The BBC story is a useful warning.
AI is changing citizen behaviour.
The volume of information coming into public bodies may increase. The complexity of that information may increase. And councils cannot simply respond by asking already-stretched teams to read more emails, process more pages and work longer hours.
The answer is not to fight the citizen using AI. It is to fight the administrative noise using AI.
Copilot Agents provide a way to start doing exactly that.
Not one giant AI system trying to run the council. But a collection of carefully designed agents, each focused on a specific problem, connected to trusted information and governed appropriately.
The FOI Agent is one example of what that can look like; it points towards something much bigger: A local government workforce where AI handles more of the information burden, so people can spend more time on the decisions, conversations and services that actually require people.
AI created the noise. Maybe AI can help councils find the signal.
CPS helps local government organisations identify where Copilot Agents can create practical capacity. From FOI and complaints through to customer services, case management and wider service transformation.
We help councils move from “What can AI do?” to a much more useful question:
“Where could an agent take work away from our people — safely, securely and with measurable impact?”
Explore more insights
How AI, Copilot and Agents can help councils reduce operational friction, improve productivity and build smarter, more resilient organisations during Local Government Reorganisation.
Microsoft is retiring its twice-yearly Release Wave model and moving Dynamics 365, Power Platform and Dataverse updates into the new AI at Work Roadmap. Here’s what continuous roadmap visibility means for planning, governance and turning Microsoft innovation into business value.
Most organisations are already using AI to improve productivity. The next step is Agentic CRM, where AI can take action, make recommendations and help drive customer outcomes. Join this webinar to explore how organisations are using Microsoft Dynamics 365, Copilot and AI Agents to transform customer engagement and prepare for the future of CRM