Amazon is working on an initiative called World Product Graph, according to an internal document reviewed by Reuters. The project aims to collect data on nearly every product sold worldwide, including items Amazon does not stock.

The plan was also described by people familiar with the matter. Amazon says the document is an outdated draft and may not reflect its current plans.

What the Plan Aims to Do

The goal is to give shoppers deeper product information without sending them elsewhere. If it works, Amazon could widen its role in product research. That would put it in closer competition with Google. It could also help Amazon respond to the growing e-commerce business at Walmart.

The effort also relates to the rise of AI agents. These tools can complete shopping tasks on a user’s behalf.

The document argues that Amazon cannot serve every shopping need if customers must leave to research. In its view, answers about how products relate to each other sit on Google, brand sites, review sites and tech forums. Finding them has meant slow research outside Amazon.

Where Shoppers Start Their Search

A 2025 Morgan Stanley report shows a split in shopper behavior. Among online shoppers researching a product, 16% start on Amazon. Google draws 57%.

The picture changes when buyers already know the product. In that case, 41% of searches begin on Amazon. Google gets 23% and Walmart gets 18%.

How the Data Would Be Collected

The document says the project depends on learning how brands organize products on their own websites. Some brands and retailers would send product data to Amazon on an ongoing basis. Amazon would also crawl other retailers’ sites on its own.

Reuters could not determine how Amazon would run the crawling. It also could not confirm whether Amazon would ask permission in every case.

Amazon has reportedly signed up Coca-Cola and Bayer, among others, as data contributors. It also held a “CEO-level P&G/Amazon summit” for a trial with Procter & Gamble. None of the companies responded to requests for comment.

Better Answers and Lower Costs

With fuller data, Amazon could help users confirm the latest model or color of items such as a vacuum cleaner or tablet. It could also guide them toward options that suit them, using their shopping history and other data.

The document also points to weaknesses in Rufus, Amazon’s AI search tool. The company began calling it “Alexa for Shopping” in May. According to the document, many Rufus answers are wrong or incomplete. It adds that customers often lack the expertise to spot the errors.

The document says Rufus handles about 83 million searches each week, and that running them is expensive. More data would let Amazon use cheaper models more efficiently.

The memo includes an imagined customer quote: “It’s like the store finally knows what it’s selling.”

Amazon responded that the usage figures are significantly outdated. It says Alexa for Shopping is roughly 97% accurate worldwide.

The Pressure From AI Shopping Agents

People familiar with the matter say the project could also help Amazon counter rival AI shopping agents. Amazon executives are reportedly concerned that agents may send traffic to sites with more complete data.

CEO Andy Jassy addressed the topic on the company’s third-quarter earnings call. He said few customers shop this way yet. He added that Amazon’s advertising business has seen minimal impact. Jassy also said more than 400 million people have used Alexa for Shopping in the past year.

Meta’s Muse agent shows why the topic matters. It has attracted more than 10 million users in the US and Canada since launching in September. Amazon blocked Muse from its retail site that same month.

JPMorgan said in a research note that Muse could become the most widely used consumer AI app since ChatGPT. Elon Musk, replying to a post about xAI’s Grok being blocked on Amazon, wrote on X last month that perhaps a competitor to Amazon should be started.

According to the people, a large product data store could push AI agents toward Amazon.com instead of other retailers. That would also help protect Amazon’s growing advertising business.

Building on Existing Services

Employees describe the initiative as an early step toward a one-stop shop for nearly any product. It would expand “Buy for Me” and “Shop Direct.” Those services list items sold on other sites and let customers buy them through Amazon.

Robert Tekiela, a vice president in Amazon’s Selection & Catalog Systems unit, is organizing the project. In the document, he says shoppers have long been able to find products on Amazon. What has been missing, he says, is an understanding of how products relate, which one fits a customer, and what else to consider. Tekiela did not respond to a request for comment. One of the people said Amazon is moving staff internally to start the work.

What Remains Uncertain

The document is a PR/FAQ, a format Amazon uses for a press release and frequently asked questions. These are typically six-page memos that set out the rationale and business case for a new product. Not all of them become real services.

The document lists an aspirational release date of October 1, 2027. Reuters could not determine whether that date is realistic or whether the plan will launch. The people said the project could be changed or dropped because of cost or other concerns.

Amazon said brainstorming documents like this one do not necessarily reflect current plans. It added that it is always working on ways to make its product catalog comprehensive, accurate and helpful.

Final Thoughts

For now, World Product Graph is an idea on paper, not a confirmed product. Still, it shows how Amazon is thinking about product data as shopping moves toward AI agents. Whether World Product Graph reaches shoppers will depend on funding, execution and how brands and rivals respond.

Related post

View all