Principles and framework
RAD principles, risk model, and the decision frameworks for responsible agentic UX.
Explore →v1.0 34 Components 7 Interaction Patterns
RAD is a design system for responsible agentic experiences — with reusable patterns, components, and guidance for disclosure, oversight, recovery, and accountability.
Everything you need to design and build responsible agentic interfaces.
RAD principles, risk model, and the decision frameworks for responsible agentic UX.
Explore →Reusable flows for disclosure, approval, intervention, recovery, and auditability.
Explore →34 components that make autonomous behavior visible, controllable, and accountable.
Explore →Design assets, contribution guidelines, changelog, and implementation references.
Explore →Start with the problem
Find the right RAD pattern for your design challenge.
Users don't know what the agent is doing or why decisions were made.
Autonomous actions happen without meaningful oversight or clear intervention points.
Users can't understand why the AI made a decision or what evidence it used.
No audit trail exists for what the agent did, when, or why.
Core building blocks for transparent, controllable agentic interfaces.
Requires explicit human confirmation before consequential agent actions.
Details →Design Principles
"RAD components are not suggestions. They are design contracts — the minimum viable accountability layer for any agentic system."
— RAD Design System · Principle 1: Design for Oversight
Read the principles →