Amazon's World Product Graph Aims to Own All Product Data by 2027

Amazon has set an aspirational launch date of October 1, 2027 for a project designed to catalogue data on virtually every product sold on Earth, whether or not Amazon stocks it, according to an internal document reviewed by Reuters. The initiative, called World Product Graph, sits inside the company's Selection & Catalog Systems unit and reports to division vice president Robert Tekiela. The core mechanism is simple: if Amazon holds the richest, most accurate product data anywhere, shoppers — and the AI agents shopping on their behalf — never need to leave its site.

What does the plan actually involve?

The document is what Amazon internally calls a PR/FAQ — a roughly six-page narrative memo laying out the rationale and business case for a new product. Not all such memos become real services, and people familiar with the matter said World Product Graph could still be modified or abandoned over financial or other concerns.

The plan "depends on discovering how brands organize their products on their own websites," the document states. That means two data streams:

• Brands and retailers contributing product data to Amazon continuously
• Amazon independently crawling other retailers' websites

Reuters could not determine how the crawling would work, or whether Amazon would need permission in every case. Some groundwork is already laid: Coca-Cola and Bayer have signed up to contribute data, and Amazon held a "CEO-level P&G/Amazon summit" for a trial with Procter & Gamble. None of those companies responded to requests for comment. Amazon is also shuffling internal staff to get the project moving.

In short, the catalog becomes the moat.

Why does Amazon need more data?

Two reasons stand out. The first is search competition. A 2025 Morgan Stanley research report found that 57 percent of online shoppers start product research on Google, against 16 percent on Amazon. When shoppers already know exactly what they want, the picture flips: 41 percent of those searches start on Amazon, versus 23 percent for Google and 18 percent for Walmart.

"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 says. "These answers exist, but the only way to find them has been to leave Amazon for painstaking research."

The second reason is cost and accuracy. Amazon's AI search engine, Rufus — rebranded as "Alexa for Shopping" in May — handles about 83 million searches weekly, and the document calls those queries "expensive" to run. "Many Rufus answers are factually wrong or incomplete, but customers lack expertise to detect these errors," it states. Richer data would let Amazon run cheaper AI models more efficiently. Amazon disputed the figures, saying in a statement that "the usage figures are significantly outdated" and that its data shows Alexa for Shopping is roughly 97 percent accurate globally.

The takeaway: better data cuts both the error rate and the compute bill.

How does this connect to AI shopping agents?

The competitive pressure is real. Meta's Muse agent has drawn more than 10 million users in the US and Canada since launching in September. Amazon blocked Muse from its retail site that same month, and last year sued AI firm Perplexity to prevent similar access — a case still ongoing. Elon Musk, responding last month to a post saying xAI's Grok agent was blocked on Amazon, wrote on X: "Maybe we should start a competitor to Amazon."

CEO Andy Jassy downplayed the threat on the third-quarter earnings call, saying "few customers are doing much shopping this way yet" and that Amazon's advertising business has been minimally impacted. Yet executives, according to people familiar with the matter, worry that agents with better data elsewhere could divert customer traffic. Jassy noted Alexa for Shopping has been used by over 400 million people in the past year. JPMorgan, in a research note, warned that Muse "has the potential to become the most widely used consumer AI application since ChatGPT."

What happens next?

Amazon framed the document cautiously. "Draft, outdated documents often paint an incomplete picture and the existence of a brainstorming document such as this one doesn't necessarily reflect our current plans," the company said in a statement, adding: "We're constantly working on new ideas and technologies to make Amazon's product catalog comprehensive, accurate, and helpful for customers."

Employees describe World Product Graph as an expansion of existing "Buy for Me" and "Shop Direct" services, which list products sold on other sites and let customers buy them with Amazon as intermediary. "Customers have always been able to find products on Amazon," Tekiela said in the document. "What's been missing is understanding — how do these products relate to each other, which one is right for me, and what else should I consider?"

Whether October 1, 2027 holds is anyone's guess; the strategy behind it is not.