Shopify Opens Its Catalog to Any AI Shopping Agent
Shopify will let shoppers set up AI agents to buy products for them, opening a catalog of more than a billion listings to any agent a shopper connects. The company described the expansion on October 13, 2026. The Verge
That catalog includes items sold on Shopify storefronts and elsewhere. It is not limited to one merchant checkout or a curated marketplace feed.
Until now, agent access required a partnership integration. In September, Shopify announced a partnership with Meta to let its Muse agent browse and buy items for a person. The new model removes that requirement. Shoppers can now use any AI agent without a partnership integration.
How it works
Customers turn on the feature by connecting an agent to their account on Shop. Shop is the Shopify-owned app that brings together orders, tracks packages and recommends products.
Checkout stays inside the conversation with the agent. Customers pay with one-click Shop Pay without leaving that interaction, and orders placed through the agent flow back into the Shop app. Buyers must approve each purchase made through the agent. There is no automatic payment and no background use of a card without an explicit confirmation step.
Shopify describes the feature as a personal shopper that accepts rules, such as a spending limit or a request to show only products with fast shipping. In technical terms, those rules filter the candidate products before the agent shows options, rather than sorting the list afterward.
The plumbing underneath
The change for shoppers rests on infrastructure Shopify has built for agentic commerce, meaning shopping carried out by AI agents. Shopify Catalog works as a structured data layer, a consistent product format that makes items readable to AI agents, according to company documentation. Shopify
Shopify offers a Catalog API, a software interface connected to its global product catalog. In Winter '26, Shopify rebuilt its developer platform for AI as part of this agentic commerce push. Shopify
The company has described agentic shopping infrastructure that lets AI agents pull from merchant catalogs and build carts across stores. Shopify Catalog lets AI platforms and agents search and show product information to help customers find products. Shopify
Two recent releases added related pieces. On September 28, 2026, Shopify said browser-based AI agents can now complete purchases on Shopify merchants sites. TechCrunch On October 1, 2026, Shopify introduced Canvas, a tool that lets merchants set up a Shopify store by chatting with AI. On August 5, 2026, Shopify said AI search is bringing more traffic and sales, not replacing Google.
For comparison on scale, Google's agentic checkout can complete a purchase from a feed with more than 50 billion listings interpreted with Gemini AI. Shopify's set here is smaller, at over a billion listings, but it is linked to transactions through Shop Pay and order handling in Shop.
The broader context here is market size. The global market for AI agents in ecommerce was valued at $3.6 billion in 2024 and is projected to reach $282.6 billion by 2034. The global AI marketing market is projected to reach $82.23 billion by 2030. Shopify has also launched an AI Store Builder tool that creates three store layouts with images and text from keywords, which shows how much of the store setup process it expects to handle.
The broader picture here for merchants and tool builders is control of checkout context. Search, handling of shopper rules and product selection move into a third-party agent conversation. Payment, identity and post-purchase records stay inside Shopify's systems. That split keeps Shopify's position while giving agent developers a stable target: connect to Shop, query structured catalog data, apply buyer rules, and return a Shop Pay transaction for approval.
In my view, that design will hold up better than either fully closed checkout or fully independent agent purchasing. I have covered each interface shift as discovery moved somewhere new while payment and fulfillment settled around a smaller set of trusted systems. Browsers did this in the PC era, app stores did it in mobile, and APIs did it in cloud. Agentic commerce looks similar. The agent becomes the storefront, while money movement and order records stay with the platform shoppers already trust for refunds and tracking.
The broader question here for implementation teams is what approval plus guardrails leaves unsolved. Per-purchase approval and buyer-defined limits address the clear abuse case, but they raise product questions. How agents explain why one listing was chosen, how merchants tune structured product details for agent search rather than human browsing, and how support handles an order where the agent misunderstood intent but the buyer tapped approve. None of those are blockers. They are the normal work as a new layer settles, and Shopify has made Shop the place where that settling happens.
