Amazon has a plan to index the world’s online shopping data

Judgment

Rival retailers and brands — not Amazon — hold the live decision: whether to let Amazon crawl and index their catalogs so they surface inside AI shopping answers, trading control of their own product data for reach inside a competitor's index.

Basis

Reuters reports Amazon's "World Product Graph" would stockpile data on products sold anywhere online, including items Amazon doesn't sell, served through Amazon's own site and Alexa for Shopping (formerly Rufus). The mechanism is discovery capture: if Amazon holds the largest shopping index, both shoppers and third-party AI agents route through Amazon's pages instead of Reddit or Google. Amazon already links out to outside retailers via "shop direct," so the plumbing partly exists. The stated costs are concrete: an internal document says "Many Rufus answers are factually wrong or incomplete" and that running searches through the AI is expensive; more data is supposed to improve accuracy and allow cheaper models. Feasibility hinges, per the document, on "discovering how brands organize their products on their own websites" — likely crawling. Reuters says it's unclear if or when this ships, though the document carried an October 1st, 2027 release date. Amazon told Reuters that "Draft, outdated documents often paint an incomplete picture" and that a brainstorming document doesn't necessarily reflect current plans.