Overview
How might we design voice commands so people can trust and act on AI output?
When this pattern fits
- Perfect for mobile applications, accessibility tools, and hands-free interaction scenarios where voice commands provide natural, efficient control.
When to skip or lighten it
- Dense data entry faster on keyboard.
- Open offices where speaking commands is disruptive.
- Actions that need visual confirmation of complex forms.
States
State model coming soon
Key UX elements
Key UX elements coming soon
Anti-patterns to avoid
Rigid phrase list with no natural language fallback.
Executing destructive commands with no confirmation.
Listen state with no cancel or “stop listening”.
Failed recognition with no text fallback suggestion.
How products use it
| Product | Implementation |
|---|---|
| Siri | App intents and shortcuts via spoken commands. |
| Google Assistant | Routine and app control from natural phrases. |
| Alexa | Skill commands for smart home and media. |
| Voice access (OS) | “Show numbers” and dictation for full UI control. |
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 "Voice Commands" AI interface design pattern.
Pattern Definition:Frequently asked questions
Voice commands vs voice input?
Voice input fills text fields. Voice commands trigger actions like “open settings” or “start timer”.
How show available commands?
Cheatsheet, “what can I say?”, and contextual hints on first use of voice mode.
Confirmation for risky actions?
Use voice confirmation or on-screen confirm for pay, send, delete, and bulk operations.
Offline support?
Cache core commands on device when possible; label cloud-dependent actions.