Amazon is exploring a plan to collect information on virtually every product for sale worldwide, including items it doesn't stock, Reuters reported Friday, citing an internal document and people familiar with the matter. The project, called World Product Graph, would give shoppers fewer reasons to leave Amazon for Google, brand websites, or review sites before deciding what to buy.
What Is the World Product Graph and How Does It Work?
The draft calls for brands and retailers to supply product data on an ongoing basis. Amazon would also collect information directly from other retailers' websites. Reuters could not establish how Amazon would gather that data or when it would need permission. The memo says Coca-Cola and Bayer had agreed to contribute data and that Amazon held a meeting with Procter & Gamble executives about a trial. The three companies did not respond to requests for comment.
In practice, a product graph is a structured database that links each SKU to attributes such as brand, model number, dimensions, price, availability, and user reviews. By aggregating these data points across millions of merchants, Amazon could power search, recommendation, and comparison tools that work across the entire e-commerce ecosystem, not just its own inventory.
Why Amazon Wants a Catalog Beyond Its Own Inventory
Amazon already runs two services - Shop Direct and Buy for Me - that let shoppers discover products sold by other retailers through Amazon's platform. Those services rely on limited data feeds and manual onboarding, which restrict the breadth of items shown. A comprehensive graph would let Amazon answer queries like "latest tablet model" or "best cordless vacuum" even when the item is sold only by a competitor.
According to a 2025 Morgan Stanley report cited by Reuters, 57% of online shoppers start product research on Google, compared with 16% on Amazon. When shoppers already know what they want, 41% start on Amazon, ahead of Google at 23% and Walmart at 18%. By offering richer product data earlier in the research funnel, Amazon hopes to shift a portion of that 57% toward its own site.
How the New Catalog Could Influence Amazon's Shopping AI
The draft said Amazon's AI shopping assistant, formerly called Rufus and renamed Alexa for Shopping in May, often gave wrong or incomplete answers. It also put weekly searches at about 83 million and said better product data could help Amazon run those searches using cheaper AI models. Amazon told Reuters the usage figures were outdated and said its own data put the assistant's global accuracy at about 97%. CEO Andy Jassy said more than 400 million people had used it over the past year.
Better data would allow the assistant to retrieve precise specifications, compare alternatives, and surface relevant accessories without needing to hand off the query to a search engine. In turn, Amazon could keep the interaction within its own advertising ecosystem, where it earns revenue from sponsored product placements.
Competitive Pressures from Rival AI Agents
People familiar with the project told Reuters that Amazon executives worry rival AI agents could send shoppers to retailers with better product information. A more comprehensive catalog could give those agents a reason to search Amazon first, helping protect the company's advertising business. Jassy told analysts that relatively few shoppers use agents to buy things and that the impact on Amazon's ad business has been minimal so far.
Meta's Muse has attracted more than 10 million users in the US and Canada since its September launch, Reuters reported. Amazon blocked Muse from its retail site that month and is still in a legal dispute with Perplexity over similar access. Elon Musk also floated the idea of starting an Amazon competitor after an X post said Amazon had blocked xAI's Grok agent.
Implementation Timeline and Organizational Structure
The draft lists October 1, 2027, as a target. One person told Reuters that Amazon was reassigning staff to work on the project, which the document places in its Selection & Catalog Systems unit under vice president Robert Tekiela. Reuters could not establish whether or when the project would launch. Amazon said the draft was a brainstorming document that might not reflect its current plans.
Should the timeline hold, the next steps would likely include finalizing data partnership contracts, building ingestion pipelines to scrape or API-pull data from retailer sites, and integrating the graph with Amazon's search and recommendation engines. Regulatory scrutiny may arise, especially in regions with strict data-sharing rules, but Amazon's legal teams have experience navigating such issues.
Potential Impact on Consumers, Brands, and the E-Commerce Landscape
For consumers, a unified product graph could mean faster answers, more accurate comparisons, and fewer clicks to reach a purchase. However, critics argue that consolidating product data under a single dominant platform could reduce competition and give Amazon greater leverage over pricing and visibility.
Brands that join the initiative may gain a new channel for product discovery, but they could also face higher fees for data onboarding or be required to share proprietary specifications. Smaller retailers might find the data-submission process burdensome, potentially widening the gap between large manufacturers and niche sellers.
In the broader e-commerce ecosystem, the move signals a shift from search-engine-centric product discovery to platform-centric discovery. If Amazon succeeds, other marketplaces may be forced to develop similar graphs or risk losing traffic to Amazon's more comprehensive offering.
What Remains Uncertain
Key unknowns include the exact method Amazon will use to collect third-party data, the cost structure for participating brands, and how privacy regulators will treat the aggregation of product-level information across borders. The company has described the document as exploratory, so the project could be altered, delayed, or abandoned based on internal testing or external pressure.
Watch for announcements from Amazon's Selection & Catalog Systems unit, statements from partner brands such as Coca-Cola, Bayer, and Procter & Gamble, and any filings with competition authorities that reference the World Product Graph.