Amazon is reportedly developing a massive product information database called World Product Graph, designed to cover virtually every product sold online, including items that are not available on Amazon’s marketplace. The initiative aims to bring product information from brands, retailers and other websites into a unified system, allowing shoppers to research, compare and discover products without leaving Amazon. The proposal was detailed in an internal planning document reviewed by Reuters.
The project reflects Amazon’s growing interest in artificial intelligence-powered shopping and its efforts to strengthen its position against Google, Walmart and emerging AI shopping agents. By building a comprehensive database of products across the internet, Amazon could make its shopping assistant more useful, improve product recommendations and encourage customers to complete more of their shopping journeys within its ecosystem. However, the initiative remains a reported proposal, and Amazon has cautioned that the planning document may not reflect its current strategy.
What Is Amazon’s World Product Graph?
World Product Graph is a proposed database designed to organise information about products sold across the internet, regardless of whether Amazon itself sells them.
The system would collect product information from multiple sources, including brand websites, retailer platforms and other online resources. It would then organise the information to help users understand how different products relate to one another.
For example, a customer researching a smartphone could potentially use the system to compare different manufacturers, identify the latest models, understand technical differences and discover compatible accessories.
Similarly, someone looking for a home appliance could receive recommendations based on features, specifications, price ranges and individual requirements.
The initiative goes beyond Amazon’s existing product catalogue. Instead of focusing only on products available for purchase through its marketplace, the proposed system would aim to understand the wider online retail ecosystem.
According to Reuters, the internal document describes an ambition to make Amazon a destination for product research, not simply a platform where customers complete purchases after deciding what they want elsewhere.
Amazon Wants to Keep Shoppers Inside Its Ecosystem
One of the primary objectives behind World Product Graph is to reduce the need for customers to leave Amazon when researching products.
Online shopping often involves several steps. Customers may discover a product on Amazon, search Google for independent comparisons, visit manufacturer websites for specifications and read discussions on Reddit before making a purchase.
Amazon wants to bring more of that research process onto its own platform.
The internal planning document reportedly argues that customers should not have to visit multiple websites to understand product relationships or identify the best option for their requirements.
A more comprehensive product database could help Amazon provide information about competing products, alternative models and items available through other retailers.
This could also support existing Amazon services such as Shop Direct and Buy for Me, which help customers discover and purchase products sold outside Amazon’s traditional marketplace.
The broader ambition is to make Amazon useful throughout the shopping journey, including when customers have not yet decided which product they want to buy.
How World Product Graph Could Improve Amazon’s AI Shopping Assistant
Another important motivation is improving Amazon’s artificial intelligence shopping assistant, previously known as Rufus and now called Alexa for Shopping.
The assistant helps customers search for products, compare options and receive recommendations through conversational queries.
However, the internal document reportedly acknowledged that many Rufus answers were factually incorrect or incomplete. It also described the cost of processing customer searches through the AI system as a challenge.
A larger and more structured product database could address some of these limitations.
Rather than depending entirely on general-purpose AI models to generate answers, Amazon could provide the assistant with more reliable product information.
This could help the system respond to questions such as:
- Which smartphone offers better battery life?
- What is the difference between two generations of a laptop?
- Which vacuum cleaner is suitable for a particular floor type?
- Which accessories are compatible with a specific device?
- What are the alternatives to a product that is currently unavailable?
The database could also help Amazon use less expensive AI models more efficiently by providing them with structured information.
Amazon said its Alexa for Shopping service is approximately 97% accurate globally, while describing the usage figures in the internal document as significantly outdated.
Amazon Could Collect Product Data From Other Retailers
Building a database covering products across the internet would require information from sources beyond Amazon’s own marketplace.
According to the internal document reviewed by Reuters, the project would combine product information supplied directly by brands with information collected from other retailers’ websites.
The document identifies Coca-Cola and Bayer among companies that had agreed to contribute product data. It also mentions a CEO-level meeting with Procter & Gamble for a potential trial.
Amazon could independently crawl other retailers’ websites to collect product information, although the precise technical approach and permission requirements remain unclear.
| Data source | Potential contribution |
|---|---|
| Brand websites | Product specifications, models and descriptions |
| Retailer websites | Availability and product listings |
| Manufacturer databases | Official product information |
| Existing Amazon catalogue | Product details and relationships |
| AI systems | Product classification and information organisation |
The challenge will be combining information from different websites while ensuring that product details remain accurate and up to date.
Products frequently receive new versions, updated specifications and different regional configurations. A database covering multiple retailers would need to distinguish between these variations to avoid confusing customers.
Amazon Wants to Compete With Google and AI Shopping Agents
World Product Graph could also become an important part of Amazon’s strategy to compete with Google and the growing market for AI-powered shopping agents.
Google has traditionally played an important role in helping consumers research products before visiting online retailers.
AI shopping agents introduce another competitive challenge. These tools can search for products, compare prices and potentially complete purchases on behalf of users.
If consumers increasingly rely on independent AI agents instead of visiting Amazon directly, the company could lose control over an important part of the shopping experience.
Amazon has already taken steps to restrict access from certain third-party AI agents.
The company blocked Meta’s Muse agent from its retail website and has also pursued legal action against AI startup Perplexity over its shopping agent’s access to Amazon.
A comprehensive product database could give Amazon an advantage by making its information more useful to AI systems and shoppers.
The strategy could also protect its advertising business. Product discovery and sponsored recommendations are important parts of Amazon’s retail revenue model. If AI agents increasingly control product searches, Amazon could face pressure on both customer traffic and advertising opportunities.
World Product Graph Could Launch in 2027, but Plans Remain Uncertain
The internal planning document reportedly mentions October 1, 2027, as an aspirational release date for World Product Graph.
However, this should not be interpreted as a confirmed launch announcement.
The document is described as a PR/FAQ, an internal Amazon planning format used to explain the potential customer benefits and business case for a proposed product or service.
Not every proposal documented through this process becomes an actual commercial product. According to people familiar with the initiative, World Product Graph could still be modified or abandoned because of financial, technical or other considerations.
Amazon also issued a statement cautioning against treating the document as confirmation of its current plans.
The company said draft and outdated documents can present an incomplete picture and that the existence of a brainstorming document does not necessarily reflect its current strategy.
It added that Amazon continues to work on technologies that make its product catalogue more comprehensive, accurate and helpful.
The Bigger Picture
World Product Graph represents a broader shift in competition across the technology and e-commerce industries. Companies are increasingly competing not only to sell products but also to control how consumers discover, research and compare them.
Amazon’s proposed database could strengthen its position by combining information from across the retail ecosystem with its existing marketplace, shopping assistant and recommendation systems. This could make the platform more useful even when customers are researching products that Amazon does not directly sell.
The project also highlights the importance of product information in the emerging AI shopping market. As consumers increasingly use conversational assistants and autonomous agents to make purchasing decisions, access to accurate product specifications, compatibility information and comparisons could become a competitive advantage. However, collecting and maintaining reliable information from thousands of brands and retailers will be a significant technical and operational challenge.
Looking Ahead
The next major development will be whether Amazon turns World Product Graph from an internal proposal into a publicly available service. The reported October 2027 date remains aspirational, and the company has not confirmed a commercial launch. Its ability to secure brand participation, organise external product information and maintain accurate listings will determine whether the initiative can achieve its broader ambitions.
If implemented successfully, World Product Graph could change how consumers research products online by bringing information from multiple retailers into a single shopping environment. It could also strengthen Amazon’s AI assistant and give the company another way to compete with Google and independent shopping agents. For now, however, the project remains a reported plan rather than a confirmed new Amazon product.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
