Amazon is developing a massive product information database that could transform its online marketplace into a one-stop destination for shoppers and strengthen its position against Google, Walmart and emerging artificial intelligence shopping agents.
The initiative, known internally as World Product Graph, aims to collect information on virtually every product sold worldwide, including items that Amazon does not stock, according to an internal document reviewed by Reuters and people familiar with the project.
The proposed database would provide detailed information about products, their different models, features, colours and relationships, helping shoppers compare options and identify items that best suit their needs without leaving Amazon’s website.
The plan is also designed to address a growing challenge for the e-commerce giant as consumers increasingly turn to AI-powered tools to research products and complete purchases across different websites.
According to the internal document, Amazon wants to reduce the need for customers to search Google, visit brand websites or consult online forums and review platforms to understand product differences before making purchases.
“Amazon cannot be the go-destination for every shopping mission if customers must leave to research product relationships on Google, brand websites, and review sites,” the document stated.
The company plans to obtain product information directly from manufacturers and retailers, while also exploring ways to collect data from other retailers’ websites.
The document identifies Coca-Cola and Bayer among companies that have agreed to contribute product information to the initiative. Amazon also reportedly held a meeting with Procter & Gamble executives to discuss a trial.
See also: Top U.S. Democrats Launch Congressional Investigation Into Trump-Nvidia Chip Export Scheme
However, Reuters could not establish how Amazon intends to collect information from competing retailers or whether it would require their permission.
The project is being developed within Amazon’s Selection & Catalog Systems unit under Robert Tekiela, a vice president in the division. Staff are reportedly being reassigned internally to support the initiative.
The document, which was written in the format of an internal press release and frequently asked questions, lists October 1, 2027, as an aspirational release date. Reuters, however, could not determine whether the timeline is realistic or whether the project will be launched.
Amazon cautioned that the document may not reflect its current plans, describing it as an outdated brainstorming document.
“We’re constantly working on new ideas and technologies to make Amazon’s product catalog comprehensive, accurate, and helpful for customers,” the company said in a statement.
The proposed database could also improve the performance of Amazon’s AI-powered shopping assistant, Rufus, which the company began referring to as “Alexa for Shopping” in May.
The internal document acknowledges that many Rufus responses are inaccurate or incomplete, making it difficult for customers without specialist knowledge to identify incorrect information.
It also describes the cost of running the assistant as a concern, suggesting that a richer product database could help Amazon deliver more accurate responses using less expensive AI models.
Amazon disputed the usage figures contained in the document, saying they were significantly outdated, and said its data shows that Alexa for Shopping is approximately 97 per cent accurate globally.
The initiative comes as Amazon faces growing competition from AI agents capable of independently searching for products, comparing prices and completing shopping tasks on behalf of consumers.
Companies developing such tools could redirect shoppers to rival retailers if those platforms offer more comprehensive product information than Amazon.
Meta’s Muse agent, which launched in September, has reportedly attracted more than 10 million users in the United States and Canada. Amazon blocked the tool from accessing its retail website in the same month and has also taken legal action against AI startup Perplexity over its shopping agent.
Amazon Chief Executive Andy Jassy has played down the immediate threat posed by AI shopping agents, telling investors during the company’s third-quarter earnings call that relatively few customers were doing significant shopping through such tools.
However, people familiar with the matter told Reuters that Amazon executives are increasingly concerned about the possibility of AI agents directing customers to competing platforms.
A more comprehensive product database could help Amazon position its website as a preferred source of information for those agents, potentially drawing more shopping activity to its marketplace and protecting its advertising business.
The company is also expanding services such as Buy for Me and Shop Direct, which allow customers to discover products sold by other retailers and purchase them through Amazon as an intermediary.
According to the internal document, the broader ambition is to help shoppers understand how products compare, which options suit their needs and what alternatives they should consider.

