Work & case studies · Retail & E-commerce
Personalisation plugin for an e-commerce platform
Recommendation and search-ranking models packaged as a plugin for the client's storefront.
+18%
sales from recommended products
<80 ms
recommendation latency
Challenge
Generic "bestseller" widgets and keyword search were leaving revenue on the table.
What we built
- Session-based recommendation models and a learning-to-rank layer for search.
- A low-latency serving API with feature caching, deployed as a storefront plugin.
- A/B testing framework to prove impact before full rollout.
Results
Recommended-product revenue rose by around 18% in the A/B test, and the search ranking improvements were rolled out to all traffic.
Next case study
RAG-powered support bot for a SaaS company
Ready to chart your AI course?
A free 30-minute consultation: we listen, ask hard questions and tell you honestly where AI will and won't help.