Built on intelligence developed through billions of commerce decisions, T‑Brain brings commerce-specific AI to the systems that power modern commerce.
PALO ALTO, Calif., Sept. 30, 2026 /PRNewswire-PRWeb/ -- Today at Topsort Fall Update 2026, Topsort introduced T‑Brain, the company's first Large Commerce Model (LCM). A commerce-specific foundation model, Theo is designed to understand signals from products, catalogs, queries, shoppers and transactions—and apply that intelligence across multiple commerce decisions.
T‑Brain supports use cases spanning search, retrieval, ranking, recommendations, advertising and agentic commerce experiences. It is designed to adapt to each enterprise's catalog, shopper behavior and commerce environment, providing a shared intelligence foundation for decisions across commerce systems.
T‑Brain builds on intelligence developed as Topsort powered billions of commerce decisions across more than 100 marketplaces. Topsort first applied machine learning to a central commerce decision: which sponsored product to show, balancing shopper relevance, advertiser performance and marketplace revenue. T‑Brain extends that work beyond advertising to a broader set of commerce decisions.
"AI has changed the technology landscape quickly, and commerce is entering a period of significant change as well," said Regina Ye, CEO and Co-founder of Topsort. "I expect the online store to look fundamentally different in the next 12 to 18 months. As AI reshapes commerce category by category, businesses will need more than another assistant added to existing software. They will need intelligence that understands how their business works and helps their systems make better decisions. T‑Brain is our foundation for building toward that future."
"We've been working on this for two years. Large language models are built to understand and generate language; a Large Commerce Model learns the relationships among products, shoppers, intent and outcomes to help predict what a shopper is likely to do next," said Francisco Larrain, Co-founder and CTO of Topsort. "With T‑Brain, we're bringing that commerce-specific intelligence into a shared model foundation that can support search, retrieval, ranking, advertising and agentic experiences."
In early production deployments with some Topsort customers, preliminary results show a 21% improvement in retrieval and a 26% improvement in conversion on the commerce surfaces where T‑Brain has been applied. Topsort is also demonstrating T‑Brain at Fall Update 2026.
T‑Brain is in production with select existing customers. The company is demonstrating the model at Topsort Fall Update 2026 in Boston.
Media Contact
Holly Zeng, Topsort, 1 0000000000, [email protected], https://www.topsort.com/
SOURCE Topsort
Share this article