About.

Shopping is moving into conversations with AI agents, and the hard part is no longer finding products. It is choosing well for one particular person: what suits them, what works with what they already own, what is worth waiting for. That takes taste.

We build taste as technology. Our models represent a person the way a personal shopper with a long-standing client would: as preferences with structure, that depend on the occasion and change over time. Our solutions bring that judgment to agents, retailers and publishers, and it runs Co-Shopper (opens in a new tab), our own personal shopping agent.

The work is research-led. We publish what we think a model of taste has to be able to represent, keeping established research apart from our own position, in our research notes.

The team brings more than a decade in e-commerce, shopping and human preference together with a foundation in AI research and large-scale AI systems.