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

Amazon is developing a massive database to compile information on every product sold online, known as the "World Product Graph." This project aims to make Amazon a more comprehensive and reliable source of product information compared to other platforms like Reddit or Google. By creating a vault of data on every product sold anywhere on the internet, Amazon hopes to reduce the likelihood of shoppers leaving the site to research products elsewhere.

The World Product Graph would make it more likely for consumers to end up on Amazon even if they're searching for products sold by other retailers. This initiative could also benefit AI agents, as having the largest stockpile of shopping data may make them rely more on Amazon's product index rather than other sources. This could potentially lead to shoppers using Amazon even when interacting with AI shopping assistants like Alexa for Shopping.

To enhance the accuracy and efficiency of this system, Amazon intends to use more data to train its AI models. An internal document suggests that many current answers provided by Amazon's AI, Rufus, are factually incorrect or incomplete, and running user searches with the AI is costly. By amassing more data, the system could become more precise and efficient, potentially using less expensive AI models.

However, the implementation of this ambitious project remains uncertain. An internal document hints that the project hinges on discovering how brands organize their products on their own websites, which may involve Amazon crawling the websites of other retailers. While the project's release date has been reported as October 1st, 2027, Amazon has disputed the existence of a brainstorming document, stating that such documents do not necessarily reflect the company's current plans.