Reimagining servicing for bike retailers and owners.

ROLE
Solo
TIMELINE
Nov 2025 - Jan 2026
TEAM
Solo
01 / OVERVIEW
Service cycles are unpredictable, and there's no good way to know when a customer actually needs to come in. Findstoke was built to change that, a platform where riders connect their Strava data and shops get intelligent, usage-based service recommendations.
02 / SOLUTION
The core of the product was an AI recommendation engine that predicted when a rider's bike needed servicing based on where they were riding, how much, and in what conditions. Shops would benefit from predictable bookings, riders would get timely reminders without thinking about it. The idea came from a conversation with a shop owner while demoing something else entirely. Two months in, I hit a wall. The ride data needed to build the system wasn't legally accessible due to platform data restrictions. The product couldn't exist in the form I'd imagined.
03 / REFLECTION
I went in person to talk to bike shops before writing much code. That part I got right. What I should have done sooner was validate data access, the whole system depended on it. That lesson has shaped how I approach every project since. The technical work on the AI recommendation engine wasn't wasted though, a lot of what I learned building it came with me into Morley and Lighthouse.