Amazon is pursuing an ambitious project to catalog every product sold across the entire internet—even items the e-commerce giant does not sell itself—in an effort to keep both human shoppers and automated AI agents within its ecosystem, according to a report from Reuters.
Dubbed the “World Product Graph,” the initiative is designed to prevent buyers from migrating to search engines like Google or forums like Reddit when researching potential purchases. By amassing an exhaustive inventory of product data, Amazon also aims to position its platform as the primary index for third-party AI shopping bots navigating online retail. The company has already begun directing consumers to outside merchants through a “shop direct” feature introduced last year.
The vast catalog could also address performance hurdles facing Amazon’s built-in shopping assistant, Alexa for Shopping, formerly known as Rufus. According to an internal document reviewed by Reuters, Amazon noted that many answers generated by Rufus are “factually wrong or incomplete” and that processing user searches with the current model is costly. Feeding the system more comprehensive product data is intended to boost accuracy while enabling the use of more economical models.
Realizing the database would require Amazon to decipher how external brands organize inventory on their own websites, which may entail crawling competing storefronts. While the internal file suggested an October 1, 2027 release target, it remains unclear whether the initiative will ultimately roll out. In a statement to Reuters, Amazon downplayed the leak, saying that draft, outdated materials paint an incomplete picture and that brainstorming documents do not necessarily reflect active plans.
Source: The Verge



