Fabric builds the agent on your enriched catalog and installs it on your website, app or loyalty portal. A shopper asks one question, and the agent finds the right products, compares them, builds the basket and takes the shopper to checkout.
Shoppers describe what they want in their own words. The agent finds the matching products, explains the differences between them, and keeps the cart inside the conversation.
Shoppers can ask for everything at once. One request can include several requirements, several items, a set or a bundle, and the agent finds all of them.
The agent shows products, not only text. Product cards, side-by-side comparisons and the cart appear inside the conversation.
The agent remembers each shopper. Preferences carry through the conversation, so the agent narrows the options, builds the basket and takes the shopper to checkout.
Anthropic's Claude for Commerce blueprint is open source and a starting point. Fabric builds the shopping agent from that blueprint, connects the agent to the Fabric platform and your systems, and grounds the agent in your enriched product data, so the agent has something specific to say about every product you sell.
An agent filters before it reads. A request like "trail running, wide fit, under $150" is three requirements, and each one has to match an attribute you carry. Fabric fills the attributes that decide each category, such as width, drop and upper material for trail shoes.
The agent explains its recommendations. Context such as best use, terrain and occasion comes from your enriched product data, so the agent can say why one product fits better than another.
The catalog improves with use. Fabric refines the catalog behind the agent from real shopper interactions and from feedback across AI shopping channels.
Run the questions your shoppers ask against the agent inside Fabric. Every missed match points to a missing attribute, FAQ or piece of context that you can fix before launch.
Ask the agent what your shoppers ask. See which products the agent finds and which it misses, and fix the product data behind each miss before the agent goes live.
Set the agent's brand, tone, personality and capabilities. The agent works as an extension of your merchandising, in your voice, with the capabilities you choose.
Launch on your website, app or loyalty portal. Shopify stores use the Fabric app, and everything else connects through the Fabric MCP or API. Fabric's team installs the agent for you.
“Clone the blueprint onto a thin catalog and you have built a very articulate way to lose a sale.”
Your team works in Fabric to onboard products, enrich them, fill gaps from supplier data, publish to every channel and read the results.
Fabric is deploying the Storefront shopping agent with a small group of trial customers on top of their enriched catalogs.
Request a trial
Merchandising and analytics for AI shopping.
