Building products got easy for a while. There were patterns, playbooks, industry research, and design principles that reliably worked. For AI features, most of that doesn't exist yet.

‍

This playbook is our attempt at writing some of it down – what 19 years in UX taught us, and what we've had to figure out from scratch. You'll hear it directly from the people doing the work: our Head of Product Innovation, product managers, AI engineers, and the operations team who keep it all moving.

‍

What's inside:
‍

• Six practices from 19 years in UX: what's held up, from high-ownership teams to shipping experiments before they're perfect
• How we used our own research platform to build our own tool: the user research behind Optimal Interviews, and how it changed the product direction
• How to test AI features past the wow factor: why we run diary studies on AI features instead of concept tests
• The uncomfortable parts: why building AI features can feel like going backwards