refero-design

To design UI patterns using real-world product screens, implement a research-first methodology that searches a vast library of actual user flows to extract professional design patterns, preventing generic AI-generated layouts. This approach grounds interface creation in proven design research rather than defaults.

Research-first product design for AI agents is a methodology that accesses extensive databases of real-world product screens and user flows before generating interfaces. It guides systematic research, pattern extraction, and the application of craft principles to ensure UIs are polished and human-centered.

To avoid generic color palettes and uninspired layouts in AI-generated UI design, apply explicit anti-slop guidance rules alongside detailed craft knowledge. This involves researching successful real-world patterns and systematically applying principles for typography, color, spacing, motion, and icons.

Yes, you can use this methodology to research and design a user onboarding flow by searching over 150,000 real-world screens. The agent first researches existing onboarding patterns, analyzes successful components, and then applies craft principles to create an effective design.

Yes, this UI research approach provides access to detailed craft knowledge guides covering typography, color, spacing, motion, and icon usage. These guides ensure high-quality execution and help apply professional design craft principles extracted from real-world product screens.