Overview
How might we design live presence so people can trust and act on AI output?
When this pattern fits
- Ideal for collaborative docs, team chat with shared threads, and enterprise assistants where multiple people iterate on the same AI-generated artifact.
When to skip or lighten it
- Solo chats with no sharing model.
- Privacy-sensitive sessions where presence would leak activity.
- Async docs where presence adds noise without concurrent editing.
States
State model coming soon
Key UX elements
Key UX elements coming soon
Anti-patterns to avoid
Ghost presence after users leave.
No distinction between human and agent actors.
Presence without a way to follow someone’s view or end sharing.
Exposing presence to people outside the share ACL.
How products use it
| Product | Implementation |
|---|---|
| Figma | Multiplayer avatars and cursors on the canvas. |
| Notion | Live collaborators on shared pages. |
| Google Docs | Active user stack and cursors while editing. |
| Linear | Presence indicators on shared issues and docs. |
Implementation
Copy this prompt to generate a production-ready implementation in Cursor, Claude Code, Lovable, or any AI coding agent.
Generate a production-ready implementation of the "Live Presence" AI interface design pattern.
Pattern Definition:Frequently asked questions
Why show presence in AI sessions?
So co-editors know who is prompting, reviewing, or watching—and avoid stomping each other’s runs.
Should agents appear in presence?
Yes when an agent is actively running in the session. Label them as agents, not as fake humans.
How does presence relate to shared session links?
Shared links create access. Presence shows who is currently using that access.
What about idle users?
Dim or mark idle after a timeout so the stack stays truthful.