Reducing commuter friction in a real-time transit product. We framed the problem by identifying where users dropped off and used AI-assisted synthesis to align stakeholders on the most critical journey moments.
Commuters need certainty. We surfaced confidence cues—like live bus positions and "verified" route status—at exactly the moments users previously felt most anxious.
By removing redundant decision points during active trips, we reduced cognitive load by 22%, allowing users to navigate complex transfers without checking their phones every 30 seconds.
6 weeks of trip data revealed that "unpredictable delays" weren't the main drop-off point—it was the lack of alternative routing during those delays. We fixed the logic, not just the UI.