Notion: an agent now fires on its own the moment your meeting notes are done

What happened: On July 31, 2026, Notion added a new trigger to its Custom Agents called “Meeting note summarized.” You can set an agent to fire the exact moment AI Meeting Notes finishes writing up and summarizing a meeting — not when the page is first created, but once the transcript is actually done. That removes a chunk of manual follow-up work: updating a status tracker, posting a recap to Slack or by email, and filing tickets from feedback raised in the meeting. Notion says this trigger is more reliable than earlier approaches built on polling or on firing right at page creation — the agent waits until the AI note is genuinely finished, so it never works from an incomplete transcript and needs no manual schedule.

What it means for you: If you already use AI Meeting Notes in Notion, this fixes exactly the annoying half hour after a meeting where outputs get manually copied into other tools. Typical uses:

• Updating a project tracker — the agent fills in the latest status in a shared table or database after every meeting.
• Sending a recap — decisions and next steps from the meeting go straight to Slack or email, to the right people.
• Filing tickets — feedback raised in the meeting turns into concrete tickets for your team without manual copying.

A typical example: after a weekly team meeting, the agent writes down who committed to what directly in your tracking doc and sends a short recap to the manager — you just open the result to check it, instead of writing it from scratch. For smaller teams or freelancers who run client calls, the same idea works for automatically creating an invoice or a next-step reminder right after the call.

How to start:

1. Open the settings of an existing Custom Agent in Notion, or create a new one.
2. In the triggers section, add the “Meeting note summarized” trigger.
3. Choose which meetings, or which type of notes, should fire the agent.
4. Set what the agent should do — update a database, send a message, or create a task.

What to watch out for: You're letting an agent act automatically based on an AI summary of a meeting — before you point it at shared trackers or team communication, manually check its output for a few weeks. And if meetings cover sensitive matters (salaries, personal reviews, internal figures), consider whether the agent should have access to the full transcript or only to selected parts of the note.