Full product redesign: from outdated UI to a live App Store product with AI-generated effects, video creation, and an in-app coin economy.
Vizu is an AI photo and video maker — turn a photo into a styled image or short video using generative effects. I came onto the project without a PM and owned everything end-to-end: full product redesign from outdated UI to launch, produced all AI effect previews myself (Runway and other generative tools), designed the coin economy from scratch (earning, spending, gating, IAP), and delivered handoff. Two months from kickoff to App Store. Launched live on iOS and Web — early signal at 4.3★ / 45 reviews at handoff.
The original app had strong AI capabilities locked behind a UI that didn't communicate them — generic interface, no monetization structure, no clear path from 'try it' to 'pay for it.' With no PM and a hard launch timeline, the design also had to drive product decisions: feature scope, monetization model, and how the AI effects would be sold.
Redesign the full experience, build a coin economy that monetizes premium AI effects without breaking the core try-it-yourself loop, and ship to App Store in two months. The coin model: users earn a small starter balance, spend coins on premium effects, and buy more via IAP when they run out. Coins gate the best effects but never the act of creating — keeping the wow moment free while making depth a purchase. Two metrics drove the design:
% of users who go from trying a free effect to making their first IAP. The starter coin balance is calibrated to let them feel the value of premium effects before running out.
Repeat IAP rate and average coin spend per active user. Premium effects are priced so a single buyer naturally returns to refill — the economy compounds with engagement, not just one-time conversion.
Live on App Store at launch — 4.3★ across 45 reviews. The product was designed with two metrics in mind, instrumented end-to-end at handoff:
Tracking first-IAP rate and time-to-first-purchase. Early signal: users do engage with premium effects, but sample is too small for confidence. Need ~1000 active users before A/B testing the starter balance.
If first-IAP drops below 5%, the starter balance is too generous or the coin price too high — both are tunable without redesigning.
Tracking repeat purchases and time between IAPs. The hypothesis: users who buy once buy again within 14 days. Watching for that pattern as the user base grows.
30-day refund rate +1pp max. If users feel the coin pricing is opaque or unfair, the model breaks before scale.
The redesign replaced a generic, low-trust interface with a visually compelling product that communicates AI quality at first glance. Below is the original app — rebuilt from scratch in the new direction.

The original app — rebuilt from scratch in the new direction.
The core experience starts with a simple onboarding that introduces the app's AI capabilities. Users select a photo, choose an effect, and see the transformation in real time. The flow was designed to minimize friction and get users to their first "wow moment" as fast as possible.



I designed and defined the full logic of an in-app currency system from scratch. Coins are earned through engagement and purchased via IAP — they gate access to premium AI effects. The system needed to feel fair and transparent while driving purchase conversion. Every state — balance, spending, low-coin warning, purchase confirmation — was fully designed and specced.






A selection of screens from across the app — showing the variety of UI states, interactions, and layouts designed as part of the full product redesign.












Beyond mobile, the product was extended to a desktop web interface — giving users access to the same AI effects and creation tools from a browser. The desktop layout maintains visual consistency with the mobile app while adapting the experience for a larger canvas and mouse-based interaction.
