
AI Personalization Lifts Revenue Per Visitor 47%
SpeedMVPs built a real-time AI personalization engine in 3 weeks for an e-commerce brand with 800K monthly visitors. The brand had been showing identical product recommendations to all customers, which left conversion at 2.1% with high cart abandonment. The engine combines collaborative filtering, content-based signals and LLM-generated product descriptions that adapt to each user's browsing and purchase history, and the project lifted revenue per visitor 47%.
SpeedMVPs built a real-time personalisation engine using collaborative filtering, content-based signals, and LLM-generated product descriptions that adapt to each user's browsing and purchase history.
Book a free 30-minute callProject Overview
What We Built
- SpeedMVPs built a real-time personalisation engine using collaborative filtering, content-based signals, and LLM-generated product descriptions that adapt to each user's browsing and purchase history.
The Challenge
- An e-commerce brand with 800K monthly visitors was showing identical product recommendations to all customers, leading to 2.1% conversion rates and high cart abandonment.
Our Solution
- Real-time collaborative filtering
- LLM-powered product descriptions
- A/B testing framework
- Personalised email triggers
- Inventory-aware recommendations