Retooling a design org for the AI era
I didn't tell 22 designers to "use AI." I rebuilt what they could build.
- Role
- VP of Product Design & Research
- Team
- 22 designers & researchers
- Timeframe
- 2024–2026
- Outcome
- Design team prototyping in production code; 99% retention [VERIFY]
The thesis
Every design leader in 2025 said "designers should embrace AI." Most meant a workshop and a license. I meant changing the material my team works in. A designer who can only make pictures of software must argue for their ideas. A designer who can make working software can demonstrate them.
The tooling
I built the bridge myself: a prototyping skill that lets designers generate working Angular prototypes against our real design system — real components, real tokens, real data shapes — from inside Claude Code. Not a toy sandbox; the actual material. [CLEARANCE] internal tooling detail. Then a career framework with AI fluency built in, so "builds with AI" is a promotion criterion, not a personality trait.
What changed
[VERIFY] prototype velocity, cycle-time delta, team sentiment. Designers stopped opening tickets asking engineering to "code it up" so a stakeholder could react to it. Concept conversations happen on working prototypes that are explicitly disposable — which keeps them honest. The team I retooled is also the team that stayed: 99% retention across the transition [VERIFY].
Change what designers can build, and you change what they believe is possible.