AI Search2026-11-14
The Verge
Amazon plans AI index of online products
Amazon is reportedly building a sweeping index of products sold across the internet, including items it does not sell, in an effort to strengthen AI-driven shopping search. The project, described as a World Product Graph, would collect and organize product information from many sources so that Amazon can offer richer comparisons and answers when shoppers search. Reuters reported that the goal is to keep customers on Amazon by giving them comprehensive product data rather than sending them elsewhere to research. The idea reflects a broader race to combine AI search with commerce data. Shoppers increasingly expect conversational answers, side-by-side comparisons, and recommendations that go beyond a simple list of listings. A large product graph could help Amazon's AI systems understand specifications, variants, prices, and relationships between items. It could also create new advertising and discovery opportunities, because Amazon could shape what shoppers see at the moment they are deciding what to buy. But the project raises difficult questions. Where will the data come from, and how will Amazon handle accuracy and outdated listings? Will third-party retailers and brands have any control over how their products are represented? Regulators may also ask whether a dominant platform should be able to build a comprehensive index of competing products and use it to influence shopping. For Amazon, the challenge is balancing ambition with trust. A useful product graph could make AI shopping assistants more reliable and reduce friction for consumers. A poorly governed one could amplify errors or entrench Amazon's gatekeeping power. The project shows how AI is turning product data into strategic infrastructure, and how the next phase of e-commerce competition may be fought over who organizes the world's product information.