ReFiBuy Brings Continuously Improving Product Data to Claude Commerce Agent

ReFiBuy today announced support for the Claude Commerce Agent, bringing continuous product data optimisation to shopping experiences on brands’ and retailers’ own websites and apps.

Built to work with the commerce blueprint Anthropic introduced on the 2nd of September, the connection supplies agents with enriched product data and returns context from shopper conversations to guide further catalog improvements.

Anthropic’s Claude for Commerce blueprint gives businesses a starting point for building shopping agents that help consumers describe what they need, compare products and build a cart.

As commerce teams build and test these experiences ahead of the holidays, they need product data that gives agents enough context to understand shopper needs and support purchase decisions. For more than 20 years, brands and retailers have optimised their product catalog content for keywords.

The Agentic Commerce era demands deeper context: who a product suits, how it performs, which occasions it serves and how it compares with alternatives. Without that context, an agent can overlook a suitable product because it lacks the evidence to connect it to a shopper’s needs, putting the sale at risk.

ReFiBuy addresses this through five levels of Agentic Product Enrichment, building on core product attributes with conversational attributes, shopper context, market and buyer context, and agentic context.

ReFiBuy’s bi-directional connection to Claude’s Commerce Agent allows brands and retailers to leverage their product catalog and Commerce Intelligence Engine investment in a new Agentic Commerce surface. In one direction, we automatically and continuously publish product catalog optimisations from the merchant to their Claude Commerce Agent.

In the other direction, we capture anonymised shopper behavioral context from Claude Commerce Agent and that becomes an input into the closed-loop reinforcement learning (RL) product catalog improvement cycle.

– Scot Wingo, CEO and co-founder, ReFiBuy For merchandising and catalog teams managing large assortments, this creates a practical way to prioritise enrichment based on the questions shoppers are asking. Teams can address gaps in specifications, fit details, use cases and product Q&As.

They review suggested changes before approved updates are published, making each improvement part of the catalog available to future shoppers and agents. The announcement advances ReFiBuy’s ACO Everywhere strategy: helping brands and retailers compete wherever AI agents shape buying decisions.

It brings Agentic Commerce Optimisation (ACO) into the Claude shopping experiences they build and operate themselves, connecting product discovery with continuous catalog improvement. The addition builds on ReFiBuy’s previously announced support for answer engines and Amazon’s Alexa for Shopping (AFS).

ReFiBuy and Claude Commerce Agent Demo Site To illustrate the power of deep product catalog enrichment, ReFiBuy created a demo site at www. refibuy. ai/cfc. Built on Anthropic’s Claude for Commerce demo store, the site features 500 products across 10 categories, with basic attributes drawn from brand product detail pages.

Visitors can turn ReFiBuy enrichment on and off to compare how the Claude Commerce Agent answers the same shopper questions with and without enriched product data.

Enriched product data Categories include Beauty & Personal Care; Clothing, Shoes & Jewelry; Electronics; Grocery; Health & Household; Home & Kitchen; Pet; Sports & Outdoors; Tools & Home Improvement; and Toys & Games.

The demo is open to brands, retailers and others who want to see the Claude Commerce Agent in action and explore the role of product enrichment in its recommendations. With holiday preparations underway, brands and retailers need to identify product data gaps before shoppers encounter them.

Explore the interactive demo and connect with ReFiBuy to build a product catalog ready for every AI selling surface.