The situation
Reading habit apps live or die on whether a user opens them tomorrow. Lettore came in with a clear product vision and zero infrastructure. They needed an AI layer that personalised recommendations without feeling generic, a backend that could carry the product through scale, and onboarding and retention loops that turned signups into a habit by week four.
What we did
1. AI backbone
The personalisation and recommendation engine sits at the centre of the experience. We designed the service so that recommendations get sharper with every session, surface the right book at the right moment, and never feel like a generic feed. The AI doesn’t just rank, it remembers what kind of reader you are.
2. Entire application backend
Data model, services, APIs, the lot. We treated the backend as a long-term asset, not a hack to ship. Clear separation between identity, content, reading state, and behavioural signal. Everything the team needs to ship the next ten features is there.
3. Activation & retention loops
First session: discover three books in under 60 seconds. First week: build a reading streak that the AI rewards. Month two: the habit is on autopilot. We mapped the loop end-to-end, instrumented it, and built the product around it.