AI & EXPERIMENTS
AI is a leadership problem before it’s a tooling problem. Design leaders who treat it as a novelty will fall behind; those who embed it into org strategy, product development, and team operations will compound their leverage. I’ve done this across three organizations — shipping AI-powered product features, building responsible-use frameworks, and increasing team throughput without adding headcount.
The work below spans org-level capability building, AI-powered product features, and hands-on experiments. Each initiative was tied to measurable outcomes — reduced cycle time, de-risked production features, or operational leverage that freed designers for higher-value strategic work.
Tooling, training, and guardrails that make entire design organizations more effective with AI.
What started as a scrappy personal experiment became an official team resource. Created and open-sourced a structured prompt library helping PMs and designers conduct faster competitive analysis, user research synthesis, and brief writing. Cut brief-writing time by ~40% across 3 product teams.
Built a prototype system that uses LLMs to pre-analyze design specs and surface potential usability and accessibility issues before peer review. Reduced spec review prep time by ~30%.
Designed a framework for measuring AI tool effectiveness in design workflows — comparing task completion, quality, and team satisfaction. Used to evaluate and select AI tooling for a 30-person design org, reducing tool sprawl and informing org-wide adoption decisions.
Prototyping and shipping AI features that drive measurable product outcomes.
Explored multi-agent architectures for automating research synthesis — taking raw interview transcripts and producing structured insight reports. Reduced synthesis cycle from weeks to days.
Researched and documented emerging patterns for AI-powered conversational interfaces, including error recovery, hallucination disclosure, and progressive trust. Adopted as a shared reference by product and engineering teams, reducing design iteration cycles on AI features by ~25%.
Born from the frustration of outdated component docs, this prototype generates documentation, usage guidelines, and code examples from Figma source files using AI. Eliminated ~2 hours of manual documentation per component and is now part of the team's standard publishing workflow.
Article
Lessons from growing a team of 3 to 30 while maintaining design quality.
Article
A framework for integrating AI into design organizations without hype or fear.
Talk
Keynote at DesignOps Summit 2024 on building AI-literate design teams.
Article
How product design leaders can shift the planning conversation from features to outcomes.
Article
Why DesignOps should report to design leadership, not engineering.
Start with low-risk experiments. Build team fluency through hands-on practice, not mandates.
Establish clear principles for AI use in design — including when human judgment must remain central.
Define success metrics for AI adoption. Track efficiency gains but also quality and team satisfaction.