Skip to main content
Your agent usually runs in a Waterr meeting, a room we host that you share with a link, embed or schedule. It can also join a call you don’t host: paste a Google Meet, Zoom or Microsoft Teams link and the agent joins as a participant. It runs the same scenario with the same persona, voice, knowledge and goals.

Google Meet

Joins as a guest; the host lets it in.

Zoom

Joins through the Zoom web client; the host admits it from the waiting room.

Microsoft Teams

Waits in the lobby; the organizer confirms “Admit this bot?”.
Joining Meet, Zoom and Teams is included in the Starter plan and above.

Send your agent to a call

1

Connect your calendar

Connect your calendar in Settings › Calendar.
2

Find the meeting

Your next meetings are listed on the Home page. Rows without a Meet, Zoom or Teams link say so.
3

Invite the agent

Click Invite agent and pick a scenario. The agent is scheduled, not sent straight away: it joins about a minute before the meeting starts.
4

Admit it

When the agent knocks, the host admits it like any other guest. It starts talking once it is in.
When the call ends, the transcript and analysis appear with your other sessions.

What works where

✅ Supported  ·  ⚠️ Partly, or depends on the host  ·  ❌ Not available

Conversation

On screen

Privacy and control

After the call

¹ The agent sees still frames from participants’ video, about one per second, not continuous video.
² The agent’s camera shows the persona’s photo with its name. A live talking avatar is in beta.
³ The agent posts a notice in the meeting chat when it is admitted. Some meetings don’t let guests use chat, Teams especially, so the notice may not appear. Tell participants yourself when consent matters.

Why some features are missing

In a Waterr meeting we control the whole room: the page, the screen and the recorder. In Meet, Zoom and Teams the agent is a guest in someone else’s room. It can talk, listen, see video and show a camera, but nothing else:
  • Nothing to put on screen. Present File, Screen Apps and video playback need a Waterr page to draw on. In external calls the agent isn’t given those tools. If someone asks it to show a file, it says it can’t share its screen and describes the content out loud instead. Your files still work as knowledge.
  • No recording. Recording belongs to the room. Use the meeting platform’s own recording if you need one. You still get the full transcript.
  • The host is in charge. Admitting, muting, removing and ending the meeting all follow the host’s settings. The agent can leave, but it cannot end a meeting it doesn’t own.
Running both? Write the scenario prompt so it doesn’t depend on showing slides. The same scenario then works in a Waterr meeting and in a Teams call.