Ask without interrupting the room.
Roumy / What Are Agent-Native Meetings?
Category definitionWhat is an agent-native meeting?
A meeting designed for people and the AI agents they already work with—not a video call with another recorder added to it.
Retrieve the exact detail while it matters.
Board, canvas, and voice by invitation.
The agent arrives with the person
Most meeting software begins when the call begins. A personal agent begins earlier: it already knows the person’s projects, language, decisions, and working context. An agent-native room lets that continuity enter with them.
In Roumy, the person and agent share one legible seat. The agent is not disguised as another human and does not become an anonymous third participant.
Private help comes before public contribution
A useful agent should be able to brief its person, retrieve an exact detail, or draft a response without broadcasting every exchange. The human remains in the conversation and decides what crosses into the room.
- Ask privately without interrupting the discussion.
- Retrieve context from the work the agent already knows.
- Invite the agent to speak or present only when the room needs it.
The meeting becomes an active working surface
Transcripts preserve what happened. Agent-native meetings can also shape what happens next. An agent can explain on a shared stage, build a board, update a canvas, and leave the group with an artifact it has already reviewed together.
The category in plain language.
Is an agent-native meeting the same as an AI notetaker?
No. A notetaker primarily captures the call. An agent-native meeting supports a person’s own agent before, during, and after the conversation.
Does the agent act without the person?
The product model keeps the person in control of private help, public sharing, and consequential actions.
Why does same-seat identity matter?
It makes the relationship visible: the room can see which agent belongs with which person and when that agent contributes.