Client
RetailX Global
Industry
Retail
Tech Stack
FlutterARKitARCoreFirebaseTensorFlow LiteStripe
The Challenge
What We Were Up Against
RetailX had a 32% online return rate driven by sizing uncertainty and an inability for shoppers to visualise products. Their existing native iOS/Android apps had a 2.1 store rating and were rarely updated.
Our Solution
How We Solved It
We rebuilt their mobile commerce experience in Flutter with a custom AR engine using ARKit/ARCore for real-time product overlay, a body-measurement AI for size prediction, and a unified omnichannel cart that syncs with in-store inventory.
01
UX Research
50-user study revealed sizing uncertainty and poor photography as the top two barriers to purchase.
02
Flutter Architecture
Monorepo Flutter app sharing 94% of code between iOS and Android, with platform-specific AR bindings.
03
AR Engine
Custom ARKit/ARCore integration with product 3D model pipeline — asset processing from merchant uploads automated.
04
AI Size Predictor
Users input 3 measurements; CNN model predicts size from brand-specific fit data with 91% accuracy.
05
Omnichannel Cart
Real-time inventory from 120 stores surfaced in-app — buy online, pick up in store, or return anywhere.
The Impact
Measurable Results
4.9★ App Store rating (up from 2.1★)
Return rate dropped from 32% to 21% (−35%)
220,000 downloads in first 90 days
Session duration 2.8× longer than industry average
AR try-on used in 40% of all purchase sessions
"Our customers call it 'the best shopping app they've ever used' — and they mean it. The AR try-on is something our competitors still haven't matched."
David Kim
Head of Digital, RetailX Global
Key Outcomes
App Store rating4.9★
Return rate−35%
Session duration2.8×
Downloads in 90 days220K
