Skip to content

Build a sprint-review summarizer for your Friday demo

Every Friday demo ends the same way: great discussion, then nobody writes it up. Build a no-code agent that turns the sprint-review meeting into a clean summary — what shipped, what's next, what's blocked — that you fire off in one click instead of dreading.

Stage: Agent Builder · For: Maker, End user, Developer · Level: Intermediate · Time: 20 min

When to use this

Sprint reviews generate exactly the kind of summary people want and nobody writes: demoed items, decisions, follow-ups, blockers. The inputs are consistent every week — the meeting recap, the demo notes, maybe the board. That repeatability is the signal you've outgrown one-off delegation and should build an agent: give it the format once, and turn every Friday's raw material into the same tidy write-up.

Use this when your team's sprint reviews are valuable in the room but evaporate the moment they end.

What you'll need

  • M365 Copilot license with Agent Builder (Copilot → Create an agent)
  • A consistent input the agent can ground on — your sprint-review meeting recap/transcript, a Teams channel, or a notes doc in OneDrive/SharePoint
  • A clear idea of the summary format your team wants (sections, audience, length)
  • 20 minutes and the four ingredients: name, instructions, knowledge source, starter prompts

Try it now — the build

In the Agent Builder conversation, describe the summarizer and lock the format:

Create an agent called "Sprint Recap" that turns our Friday sprint-review meeting into
a summary. Ground it on [the meeting recap / Teams channel / notes doc]. Output the
same structure every time: (1) Shipped this sprint — what was demoed, one line each;
(2) Decisions made; (3) Next sprint focus; (4) Blockers and risks, with an owner where
one was named. Keep it skimmable, plain language, no jargon. Pull only from the
meeting — if something wasn't discussed, leave it out rather than inventing it.

Why this works: it sets the knowledge boundary (the sprint review, not everything), fixes a consistent four-section format so every week is comparable, sets the tone (skimmable, plain), and adds the no-invention rule. A locked format is what makes a recurring summarizer genuinely useful — people learn where to look.

Step by step

  1. Open Agent Builder. In M365 Copilot, choose Create an agent. Build by chatting or use the Configure fields.
  2. Add the knowledge source. Point it at the meeting recap, Teams channel, or notes doc your sprint review produces each week.
  3. Write the instructions (the prompt above, adapted). The fixed four sections and "only from the meeting" rule are what keep every recap consistent and honest.
  4. Add starter prompts matching the weekly job: "Summarize this week's sprint review", "What did we commit to next sprint?", "List the blockers with owners."
  5. Test in the preview pane. Feed it a real (or recent) sprint-review recap. Check the format holds and nothing is invented — then tune the section wording to match what your team wants.

Screenshots

We deliberately don't ship screenshots that go stale — the Microsoft Copilot UI changes often. Follow the numbered steps above, which we keep current. Maintainers can regenerate fresh captures with the Playwright tool in tooling/screenshots/.

Make it better

Take it from summary to communication: - Write the share-out. "Also draft a short Teams message announcing the recap, leading with what shipped." - Tailor by audience. "Give me a one-line exec version and a detailed engineering version." - Trend it. "Compare this sprint's blockers to last sprint's — what keeps recurring?" - Share the agent. Once the format's right, share it with your team so anyone can generate the recap — it stops depending on you being there.

📚 Learn more. Browse Microsoft's Agent Library for templates, and watch April's AI Agents Academy (April Dunnam, Microsoft) for step-by-step building.

Watch out for

  • It summarizes what it can read. A thin meeting recap makes a thin summary. The better your sprint review captures notes/transcript, the better the recap — feed it the richest source you have.
  • A clean format can hide a wrong fact. A tidy summary feels authoritative; skim it for accuracy before sending, especially the decisions and owners.
  • Declarative summarizes, it doesn't trigger. This agent generates the recap when you ask — it can't automatically run the instant the demo ends. For event-triggered automation you're looking past declarative building (Stage 6+).

Where this leads (the ramp)

You built a recurring agent that produces consistent output — and immediately wished it would just fire on its own after the demo and post the recap for you. That wish (real triggers, real actions, real workflow) is exactly what Stage 6 · Copilot Studio is for.

Next: Copilot Studio → Build your first Studio agent with a knowledge source + topic

Was this page helpful?

Found an issue or want to suggest an improvement? Open a GitHub issue  ·  Star on GitHub