A cross-platform fitness companion that combines real-time biometric tracking with a personalised AI coaching layer — turning raw workout data into an adaptive training plan that evolves with the user.
The client had a first-generation fitness app built in Swift — iOS only, no analytics, no personalisation, and a 28% 30-day retention rate. Users would download, log a few workouts, and disappear. The product had strong brand recognition but the app experience wasn't converting that into habit.
They needed a cross-platform rebuild that would feel premium, add genuine intelligence, and give users a reason to come back daily.
Pulse is a React Native app that shares a single codebase across iOS and Android. The core tracking layer integrates natively with Apple HealthKit and Google Fit to pull biometric data — heart rate, sleep quality, recovery score — and feeds it into an AI coaching engine that adjusts the user's training plan in real time.
The AI coach is powered by a fine-tuned model trained on exercise science principles and the client's methodology. It doesn't just suggest workouts — it explains its reasoning, adapts based on user feedback, and learns each user's preferences over time.
Pulse launched to 4.8 stars on both stores. 30-day retention climbed from 28% to 68% — a 40% lift attributed primarily to the AI coach's daily touchpoints. The client expanded their subscription tier pricing and saw ARPU increase 3.2x within the first quarter post-launch.