Overview
How might we design multi-user awareness so people can trust and act on AI output?
When this pattern fits
- Essential for family voice assistants, shared smart displays, and office ambient agents where identifying the current speaker and scoping to their permissions is required for accuracy and safety.
When to skip or lighten it
- Single-user personal devices with one account.
- Identification confidence too low for secure scoping.
- Household profiles forbidden by product policy.
States
State model coming soon
Key UX elements
Key UX elements coming soon
Anti-patterns to avoid
Reading one user calendar to another without identification.
Voice match with no fallback when unknown speaker.
Silent switch between profiles mid-session.
Kids mode with same purchase permissions as owner.
How products use it
| Product | Implementation |
|---|---|
| Alexa | Voice profiles for personalized results and shopping. |
| Google Voice Match | Speaker ID for calendar and payments. |
| Apple Home | Multi-user recognition on HomePod for personal requests. |
| Sonos Voice Control | Household voice profiles on shared speakers. |
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 "Multi-User Awareness" AI interface design pattern.
Pattern Definition:Frequently asked questions
How identify users?
Voice print, phone proximity, explicit profile picker, or PIN for sensitive tasks.
Unknown speaker?
Generic mode with no personal data and prompt to enroll or switch profile.
Multi-user vs activation boundaries?
Boundaries mark session on or off. Multi-user picks who is inside the session.
Privacy?
Show which profile is active and let guests use limited guest mode.