Case Studies
Where AI met a real person, and what I chose to do about it. Every case study is built the same way: the decision, the trust and control states, what happens when it’s wrong, the evidence, and what it cost. Read them in order or jump to the one that matches your question.
One AI assistant for 70+ products, built from zero, with a person always one action away.
Live in 40–50% of the portfolio by July 2026, and the foundation for Ecolab’s agentic platforms.
Read the case study: Enterprise AI Assistant, End to End· 13 min read
Every AI product was out of compliance. In three weeks I designed the disclosure and consent system outside AI-law counsel approved on the first review.
Eight patterns, one component and guidance that keeps future products in compliance.
Read the case study: Responsible AI, Written Down· 12 min read
How design fits with AI: a process for 80 to 100 designers, proved with a design system generated through Claude and Figma in 60 to 80 hours.
Presented as a proof of concept, shipped for real and taught to the designers.
Read the case study: An AI-Native Design Process· 13 min read
Agents watch the operation and come to the user, so no decision falls through the cracks. Piloted with nearly a dozen client partnerships.
Insights in the chat had 76% higher completion than a deterministic dashboard.
Read the case study: Agentic AI That Notices First· 11 min read
Case Study 1 is the foundation: the whole story, start to finish, and the one the other three build on.
