Notion's AI Meeting Notes can now trigger a Custom Agent as soon as a meeting summary is complete. The new connection turns the finished note into an event that can start follow-up work, closing a gap between recording what happened and acting on it.
Until now, a meeting summary could gather decisions and tasks in one place, but somebody or another automation still had to move that information into the next system. Notion's new trigger lets a configured agent begin that hand-off without a fresh prompt.
The summary becomes a starting signal
The trigger is called Meeting note summarized. A user adds it in the Custom Agent settings and selects the meetings that should activate the workflow. Once an eligible AI Meeting Note finishes processing, the agent can use the resulting content to perform its assigned task.
Timing matters here. The agent starts after the summary is ready rather than responding to an unfinished transcript. That gives it a more structured record containing the decisions, discussion and action items that Notion's meeting feature has extracted.
The selection step also provides some scope control. A team does not have to send every conversation into the same automation. Different meetings can feed different agents, or sensitive and informal calls can be left outside the trigger entirely.
Three hand-offs Notion is targeting
Notion highlights project tracking as one use. An agent can take the agreed actions from a meeting and update the relevant tracker, reducing the lag between a decision and the project record that is meant to reflect it.
A second pattern is communication. The agent can prepare or send a recap through Slack or email, carrying decisions and action items to people who were not in the room. That can make the meeting note a shared operational record instead of a page that participants have to remember to circulate.
The third example connects feedback to engineering work. An agent can turn comments captured in a meeting into tickets, creating a more direct route from customer, design or review conversations to a development queue.
Configuration determines whether it helps
The trigger removes a manual step, but it also makes the agent's instructions more important. A vague workflow could create noisy tracker updates or distribute a recap more widely than intended. The useful version needs a clear destination, predictable fields and boundaries around which meetings qualify.
Human review may still be appropriate when a summary contains ambiguous ownership or a discussion that did not reach a firm decision. The release announcement describes the trigger and suggested workflows; it does not claim that every extracted action is correct or that the agent can resolve uncertainty in the underlying conversation.
Automation begins where the call ends
This is a relatively focused release, but it joins two parts of Notion's AI product in a practical way. Meeting Notes produces a structured account of a conversation, while Custom Agents can apply instructions across connected work. The trigger makes the output of one the input of the other.
For teams already keeping projects and meeting records in Notion, that can shorten the distance between agreement and execution. The real gain is not another generated summary; it is the ability to make a completed summary start a defined process while the meeting is still fresh.