# Council Performance Reports With AI: Clear Annual And Corporate Plan Updates

URL: https://www.starien.com/blog/council-performance-reports-with-ai-clear-annual-and-corporate-plan-updates
Published: 2026-09-25
Author: Geoffrey Moffett
Topics: performance reporting, annual report, councils, corporate plan, AI

> How councils can use AI to turn performance data, corporate plan updates and annual reports into clear, honest, resident-friendly content, with careful checks on every figure.

## Quick answer

AI can help councils turn performance reports and corporate plan updates into clear summaries for residents, members and staff by drafting plain-English narratives, highlighting key results and creating web and social versions. Every figure must be checked against the source data, and reports should present underperformance honestly as well as successes.

## Key takeaways

- Residents want to know what changed for them, not how many actions are 'on track'.
- AI turns dense tables into readable narratives quickly.
- Every figure is checked against the source, without exception.
- Honest reporting of what didn't go well builds credibility.

Councils produce a lot of performance information: corporate plan updates, annual reports, improvement reports, service dashboards. Much of it is accurate and thorough, and almost none of it is read by residents. Council performance reports can be made clear and engaging with AI help, without losing accuracy or honesty.

## Why Council Performance Reports Deserve Better

- **Accountability.** Residents have a right to understand how their council is doing.
- **Trust.** Clear, honest reporting shows confidence and openness.
- **Members' scrutiny.** Clear summaries help members focus on what matters.
- **Staff pride.** People like to see the difference their work makes.

## The Problem With Most Reports

- tables of indicators with no explanation;
- RAG ratings ("amber") that mean nothing to residents;
- jargon and acronyms;
- every achievement celebrated, every shortfall buried.

## Step 1: Start With Verified Data

AI must never be the source of a performance figure. Gather:

- verified performance data from the responsible officers;
- officer commentary explaining results;
- the corporate plan objectives they relate to.

Check your AI policy. Published or soon-to-be-published performance data is usually suitable for approved tools, but unpublished data or anything with personal information may not be.

## Step 2: Draft Narratives

> "Using only the data and commentary below, write a plain-English summary of performance for [objective] for residents. Under 150 words. Explain what the results mean for people, not just the numbers. Include both progress and areas that fell short, with the reasons given. Don't calculate new figures, round numbers differently or add claims not in the source. Mark gaps with [CHECK]."

The instruction not to calculate is important. AI tools can make arithmetic errors or derive percentages that look plausible but are wrong.

## Step 3: Check Every Figure

Use a strict routine:

- [ ] Every number matches the source data exactly
- [ ] Units and time periods correct ("per year", "since April")
- [ ] Comparisons are fair (same period, same measure)
- [ ] Claims of improvement are supported by the data
- [ ] Shortfalls reported honestly
- [ ] Second person checks key statistics

## Step 4: Be Honest About What Didn't Go Well

Readers trust reports that acknowledge problems. For each shortfall, explain:

- what happened;
- why (as stated by the service);
- what's being done.

AI can help phrase this clearly and without defensiveness:

> "Rewrite this explanation of a missed target so it's clear, honest and not defensive, in plain English. Keep the facts exactly."

## Step 5: Create Versions For Different Audiences

From the approved report, AI can help produce:

- a two-page resident summary;
- a web page with key highlights and links to the full report;
- social media posts on individual results (with real photos of the services involved);
- an internal staff version celebrating teams' contributions;
- an easy-read summary.

## Step 6: Show, Don't Just Tell

Numbers land better with stories. Pair key results with a short example: a family that used a new service, a community group that benefited from a grant (with consent). AI can help draft case study questions and edit interviews into short stories, but the stories must be real and approved by the people in them.

## Step 7: Design For Reading

- headline results first;
- simple charts, clearly labelled, with text explanations (for accessibility);
- no unexplained RAG ratings or acronyms;
- a clear web version, not just a PDF.

Clear, well-structured web versions also help AI assistants answer residents' questions about council performance accurately.

## A Made-Up Example: From Table To Story

Imagine a corporate plan update table includes the line: "Indicator 4.2: Percentage of household waste sent for reuse, recycling or composting. Target: [x]%. Actual: [y]%. Status: Amber."

For a committee, the table may be enough. For residents, it means little. Using the verified figures and the service's commentary, AI drafts: "We recycled or composted [y]% of household waste this year, short of our [x]% target. The service says [reason given in the commentary]. Over the next year, we're [action from the commentary]." An officer checks every figure and phrase against the source, and the head of service approves the wording.

The result is clear, honest and useful, and it took minutes rather than an hour.

## Make It Part Of The Reporting Cycle

Rather than adding AI as an afterthought:

- build the "resident summary" into the reporting timetable;
- agree who drafts, who checks figures and who signs off;
- keep a prompt and checklist specifically for performance narratives;
- publish summaries at the same time as the formal report, so the two always match.

Over time, services can provide commentary that's written with residents in mind from the start, which makes the whole process easier.

## Questions To Ask Of Every Draft

Before any performance narrative is approved, ask:

- Would a resident understand what this means for them?
- Is every figure traceable to verified data?
- Have we described shortfalls as clearly as successes?
- Are comparisons fair and like for like?
- Does the tone sound like the council, not a press release?

If the answer to any is no, edit before it goes out.

## Pitfalls To Avoid

- **AI-generated figures or calculations.**
- **Spin.** Overstated achievements undermine trust.
- **Burying shortfalls.**
- **Stock or AI images** in place of real services and people.
- **PDF-only reports.**

For the wider risks, see [avoiding AI pitfalls in council communications](/blog/avoiding-ai-pitfalls-in-council-communications/).

## How Starien Can Help Train Council Teams

During six years leading Coleraine BID's digital delivery, we regularly turned activity and results into clear reporting for a board and the public. Our [AI Coaching](/ai-coaching/) helps policy, performance and comms teams draft clearer reports safely, and our [Public Relations](/public-relations/) approach helps tell the story honestly. [Get in touch](/contact/) to talk about training.

## Frequently asked questions

### Can AI write a council annual report?

AI can help draft sections and summaries from verified data and officers' notes, but officers must check every figure and claim, and the report goes through normal approval.

### How do we stop AI getting performance figures wrong?

Provide the figures in the prompt, tell AI not to calculate or add numbers, check every figure against the source data and ask a second person to check key statistics.

### Should performance reports include failures?

Yes. Explaining honestly where targets weren't met, why and what's being done builds trust and supports proper scrutiny.
