Bayes Centre

Building an AI Shopping Assistant with Multimodal Machine Learning

Moonsift is building an AI-powered product research tool for taste-driven purchases. Their platform helps tens of thousands of people make considered purchasing decisions every week across all retailers.

The Partnership

Moonsift partnered with the Bayes Centre to build an intelligent agent that retrieves and refines product results based on a shopper's taste profile and natural language interactions. The project employs and develops cutting-edge multimodal machine learning models.

The Technology

The project uses multimodal AI to semantically segment users' browsing behaviour, allowing the system to understand whether specific product attributes were deliberately searched for or merely displayed together by retailers. This approach enables the platform to perform thousands of hours of product research in seconds.

About Moonsift

Moonsift's category-leading shopping tool works across all retailers and is used by personal stylists, prop buyers, and creators who run businesses powered by the technology. The platform is pioneering cross-platform personalisation where each shopper has their own product discovery AI that travels with them, rather than being siloed within individual retailers.

Academic Collaboration

By partnering with the University of Edinburgh through the Bayes Centre, Moonsift gained access to world-leading expertise in multimodal machine learning and AI. This collaboration demonstrates how cutting-edge academic research can be applied to solve real-world challenges in e-commerce and consumer technology.