Amazon is preparing a massive data‑aggregation effort called the World Product Graph, a system designed to collect detailed information on virtually every consumer product sold worldwide. The initiative, described in a six‑page internal draft, targets a potential rollout on October 1 2027, though the company cautions that the document does not guarantee a firm launch schedule. The effort is being led by Robert Tekiela, vice president of the Selection & Catalog Systems division, and is intended to enhance the accuracy and cost‑efficiency of Amazon’s AI‑driven shopping assistant, now branded Alexa for Shopping.

Scope and data sources

According to the draft, the World Product Graph will pull product data from two primary channels. First, brands and retailers can voluntarily submit structured information directly to Amazon. Second, Amazon plans to deploy software that scrapes product listings from competing e‑commerce sites, a move that raises potential copyright and permission concerns. The document cites early participation from Coca‑Cola and Bayer, which have agreed to provide their product details. No confirmation has been given on how Amazon will handle data from other websites or whether it will seek explicit consent.

Strategic motivations

The project aims to close a research gap identified by Amazon: shoppers often turn to Google for product discovery, with a 2025 Morgan Stanley report showing 57 % of online shoppers starting on Google versus only 16 % on Amazon. By offering richer, more reliable product information, Amazon hopes to capture a larger share of the early‑search market and reduce reliance on external AI agents that could direct users to rival retailers. Executives have expressed concern that tools such as Meta’s Muse, Perplexity, and xAI’s Grok could erode Amazon’s traffic by providing superior product comparisons.

Impact on Alexa for Shopping

Alexa for Shopping, formerly known as Rufus, processes an estimated 83 million searches weekly, according to the internal memo, though Amazon later described those figures as “significantly outdated.” The company claims the assistant now delivers correct answers 97 % of the time worldwide. By integrating the World Product Graph, Amazon expects to rely on cheaper AI models while maintaining answer quality, potentially lowering operating costs for the service.

What to watch next

Analysts will monitor whether Amazon proceeds with the October 2027 launch timeline and how it navigates legal challenges around web‑scraping. Further partnership announcements beyond Coca‑Cola and Bayer could signal broader industry buy‑in. Finally, the competitive response from AI agents—particularly Meta’s Muse, which recently surpassed 10 million U.S. users—will be a key indicator of whether Amazon’s data strategy can blunt the “agent threat” to its e‑commerce dominance.