Mirai
A mobile wellness companion that grew into a daily habit
Mirai's launch app had strong onboarding but collapsing D30 retention. The team suspected the daily check-in flow was too long, but didn't have the data pipeline to prove it or the mobile bandwidth to iterate.
Rebuilt the check-in flow in SwiftUI + Jetpack Compose with an experimentation layer, shipped an event pipeline into BigQuery, and added a Claude-powered summary feature that turned a week of journaling into a single coaching insight. Three-month embedded pod.
D30 retention moved from 14% to 37% over two releases. The AI summary feature became the most-used weekly surface, and the team picked up the experimentation discipline to run the next iterations without us.
“We came for a mobile rebuild and left with an experimentation habit. That's the part we couldn't have bought on its own.”
Have a similar challenge?
Talk to the practice lead who shipped this one. Replies within one business day.
Talk to an expert