Per-Action Autonomy
Per-action autonomy is an AI interface design pattern that sets an agent's level of independence at the capability level rather than the app level, so different actions inside the same agent can have different approval requirements. This UX pattern lets users say 'auto-send drafts to known contacts but require approval for new ones' or 'auto-commit to feature branches but never to main,' matching the gradient of real-world risk instead of forcing a single all-or-nothing autonomy choice. Each capability carries its own autonomy policy, visible in the permissions surface and editable without unwinding the rest of the integration. The pattern is how mature agentic products move past the crude ask-every-time versus do-everything binary, giving users calibrated control where trust and risk actually diverge.
Use Case
Essential for AI coding agents, workflow automation, and productivity assistants where different actions carry different risk profiles and users need calibrated autonomy per capability, not per app.
Examples in Wild
Use this pattern in your project
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 "Per-Action Autonomy" AI interface design pattern.
Pattern Description:Send customer email
Push to remote
Post to #general
Saved per action type. New runs pick up these defaults.
Related Patterns
Get new patterns by email
Weekly AI interface UX notes and resources on Substack, no spam, unsubscribe anytime.