Work

Designing for a new reader.

At Tailor, I lead product design for complex ERP workflows. As AI agents build more of the interface, I’m making our design decisions easier for them to understand and apply—from page structure to interaction behavior.

The system before the UI.

I start with the user’s task, then work with engineering to define reusable layouts, components, and patterns. The goal is to give agents a clear foundation, so each new screen follows shared decisions.

Give each component context.

Each component needs more than a list of props. Its Markdown reference explains its purpose, when to use it, and how it behaves.

I shape the design guidance; engineering connects it to APIs and working examples. Together, these give agents the context to use the component well.

Connect patterns through skills.

Patterns explain how components work together: a table for scanning records, a modal for a quick edit, or a wizard for a longer task.

The AppShell skill guides agents to the relevant patterns and component references, turning design knowledge into a practical building workflow.

Keep the system learning.

When a component or pattern changes, its guidance and examples change with it. We regenerate the documentation and skill, then run checks for outdated references and incompatible APIs.

My focus is making this update loop part of everyday development, so agents build from current guidance.

A better foundation for everyone.

For our team, shared decisions mean less repeated work and more room to explore. For developers, they offer a reliable starting point. For the people using the software, they help create familiar, consistent interactions.