I diagnose the gap between what a system assumes about people and how they actually behave, then redesign the service around the truth rather than the assumption.
People kept dropping out of a chronic care program and it looked random. It was not. I found six mindsets underneath the drop offs and turned a lot of firefighting into something the team could see coming.
The sign up flow was great at taking payments and bad at picking people who would actually succeed. I moved the gate from can you pay to are you ready and the escalations that used to pile up afterwards mostly went away.
There was a summary panel built for the care teams and they quietly ignored it, running the whole clinic off a hidden spreadsheet instead. I dug into why and rebuilt the information architecture so the panel finally matched how they think.
Whether the users are a care team inside the business or the customers it serves, I am drawn to the problems where the real issue sits between people, tools and the way everyone assumed it would go.
Say hello → See the work