Skip to content

Answer common questions with Agents in Communities

Drop a shared agent into a Viva Engage community so the most-asked questions answer themselves — where a broad audience already gathers, instead of in a dozen private DMs.

Stage: First-Party Agents · For: Champion, IT/Admin · Level: Starter · Time: 10 min

Status: Public preview — verify current availability on the Agents in Microsoft 365 roster.

When to use this

The same five questions keep landing in your community feed — "what's our policy on X," "where's the onboarding doc," "who owns Y." Agents in Communities lets you place a shared agent inside a Viva Engage community so members get an answer from approved sources right where they asked, and everyone else sees it too. It's the broadest-reach version of self-serve: one good answer, surfaced to a whole community instead of repeated one-to-one.

This is a champion-and-admin move — you're not building an agent for yourself, you're putting a shared one where your people already are.

What you'll need

  • M365 Copilot license and a Viva Engage community you steward or admin
  • A knowledge source the agent can answer from — an FAQ, a policy site, an onboarding hub
  • A short list of the questions your community actually repeats (your "answer these" starting set)

Try it now — the prompt

Once a shared agent is available in the community, members invoke it inline:

(In a Viva Engage community) @[agent] what's our current policy on [topic], and
where's the source doc so I can read the detail?

Why this works: it asks for both the answer and the source link, which is the pattern that makes a community agent trustworthy — people get the quick answer and a path to verify it, so the agent builds credibility instead of becoming another thing nobody believes.

Step by step

  1. Pick the community and the knowledge it should answer from. Choose a community with real recurring questions and a clean, approved source to ground on — garbage in, garbage out.
  2. Make the shared agent available in the community. Add it so members can @mention it inline, the same way they'd tag a person.
  3. Seed it with the real questions. Post the handful of questions your community repeats so early answers are visibly useful — the first good answer is what earns the next hundred.
  4. Point newcomers at it: 
    Draft a short pinned post welcoming new members and showing them how to ask
the community agent a question, with two example prompts.

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

A community agent is only as good as the loop around it: - Watch what it can't answer. The questions it misses are your content backlog — each gap is a doc to write or a source to add. - Pair it with Agents in Channels. Communities reach broad audiences; channels serve focused teams. The same knowledge source can power both, meeting people wherever they already gather. - Keep the source fresh. A community agent answering from a stale policy doc is worse than no agent. Assign an owner to keep the grounding current.

📚 Learn more. The Agents in Microsoft 365 hub describes the shared-agent surfaces in plain language, and Nicole Herskowitz's (CVP, M365 Copilot) blog on enabling human-agent teams covers how shared agents work alongside people.

Watch out for

  • Public answers raise the bar. A wrong answer in a community is wrong in front of everyone. Ground the agent tightly and tell members to verify against the linked source.
  • Scope it to what it knows. An agent that confidently answers outside its knowledge erodes trust fast. Keep its grounding narrow and its "I don't know, here's who to ask" behavior explicit.
  • Stewardship is ongoing, not one-time. Standing up the agent is the easy 20%. Curating its sources and watching its misses is the 80% that makes it actually useful.

Where this leads (the ramp)

Sharing a prebuilt agent with a whole community is the last step before you start making your own. Once you've felt how much repeat work a well-grounded shared agent absorbs, the natural next question is "what if I built one tuned exactly to my team's docs?" That's Stage 4 · Agent Builder: your first declarative agent, grounded on a source you choose.

Next: Agent Builder → Build a "team knowledge" agent over a SharePoint site

Was this page helpful?

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