On November 13, 2026, Reuters reported, citing an internal Amazon document, that Amazon is advancing a project called World Product Graph: an index of product data for everything sold anywhere on the internet — even items Amazon does not sell itself — to be served through its own website. The aim is stated plainly: keep shoppers doing product research on Amazon instead of leaving for Reddit or Google. The same goes for AI agents: whoever holds the most complete product data is the index agents are most likely to use first.
Fixing its own shopping assistant first
The immediate driver sits with Amazon's own AI shopping assistant. The internal document says many of Rufus's answers are "factually wrong or incomplete," and that running user searches with AI is expensive. More data is meant to make answers more accurate while allowing cheaper models. Rufus has since been renamed Alexa for Shopping, and the shop direct feature Amazon launched last year already surfaces links to outside retailers' products on its site — World Product Graph looks like the next step on the same road.
How it would work, and where it stalls
On method, the document says the project "depends on discovering how brands organize their products on their own websites," which could mean Amazon crawling other retailers' sites. The document carries an October 1, 2027 target date, but Reuters stresses it is unclear whether or when the plan will be implemented. Amazon's reply to Reuters left room too: "Draft, outdated documents often paint an incomplete picture and the existence of a brainstorming document such as this one doesn't necessarily reflect our current plans." For readers, that means this is still an idea inside an internal document, not a product already being built. Amazon has been cautious in both directions on agent commerce: it gives third-party sellers an AI agent to watch prices and manage stock, while it blocked Meta's Muse from buying on its behalf. Three things are worth watching next: whether that date holds, whether brands agree to share their data, and whether crawling rivals' sites draws pushback from regulators and competitors.
