Amazon users run about 83 million searches a week on its Rufus shopping assistant, according to an internal Amazon document Reuters reported on 13 November, and the document calls running them “expensive”. Amazon told Reuters that “the usage figures are significantly outdated”. The document lays out World Product Graph, a plan to gather data on virtually every item for sale, whether Amazon stocks it or not.
The people familiar with the matter told Reuters the plan would also help stave off AI shopping agents. Amazon blocked Meta’s Muse agent from its retail site in September, Reuters reported. On Forter’s agent order data, Muse bought at 74% of stores with any agentic order against 24% for OpenAI’s checkout, about 3.1 times the reach by our arithmetic.
Business Pill · Agentic Commerce Explained
An evergreen Business Engineer explainer on agentic commerce, when the shopper is an AI agent. It teaches the concept and does not discuss Amazon’s plan.
Go deeper: who owns product data for AI shopping agents. Six layers sit between a brand’s data and an agent’s checkout, from Amazon’s catalog and Google’s Shopping Graph of more than 50 billion listings to OpenAI’s checkout protocol and the card networks: Amazon vs Google vs OpenAI: Who Owns Product Data for Agents.
The key insight: An internal Amazon document reviewed by Reuters puts Rufus at about 83 million searches a week and plans World Product Graph, data on virtually every item for sale; Amazon told Reuters the usage figures are significantly outdated and that Alexa for Shopping is roughly 97% accurate globally.
Reuters reviewed the document and spoke to people familiar with the matter. The document says: “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”. It is a PR/FAQ, Amazon’s six-page narrative memo format, and it carries an aspirational future release date of 1 October 2027, Reuters reported. Reuters could not determine if or when the plan may be implemented.
World Product Graph “depends on discovering how brands organize their products on their own websites”, the document says. Some brands and retailers would contribute data continuously, and Amazon would also crawl other retailers’ websites independently. Reuters could not determine how Amazon planned to crawl the sites or whether it would need to seek permission in all cases.
Amazon has already signed up Coca-Cola and Bayer, among others, to contribute product data, the document says, and held a “CEO-level P&G/Amazon summit” for a trial with Procter & Gamble. The companies did not respond to Reuters’ requests for comment. The project sits in Amazon’s Selection & Catalog Systems unit under vice president Robert Tekiela, who did not respond either.
Where online shoppers start: Morgan Stanley 2025 research report figures as reported by Reuters on 13 November 2026. The 3.6 times ratio is our arithmetic.
What Was Posted on X
JUST IN: Amazon to retrieve data on nearly all products sold on earth to compete with Google
Kalshi, the prediction-market company, posted this at 12:15 UTC on 13 November, about an hour after Reuters published its report, summing the story up as Amazon retrieving data on nearly all products sold on earth to compete with Google. The post adds no figures or sourcing of its own; this piece takes the story from Reuters, not from the post.
Rufus, by Amazon’s Own Document
The document says: “Many Rufus answers are factually wrong or incomplete, but customers lack expertise to detect these errors”. More complete data would let Amazon use less expensive models, more efficiently, Reuters reported from the document. Amazon began calling Rufus “Alexa for Shopping” in May.
In its statement to Reuters, Amazon said “the usage figures are significantly outdated” and that its data shows Alexa for Shopping is roughly 97% accurate globally. Amazon also said: “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”. Andy Jassy said Alexa for Shopping has been used by over 400 million people in the past year, Reuters reported.
Where Shoppers Start, and Who Amazon Blocks
Reuters cited a 2025 Morgan Stanley research report: 16% of online shoppers begin researching products on Amazon, against 57% on Google, about 3.6 times as many by our arithmetic. When customers know exactly what they plan to buy, 41% of searches start on Amazon, against 23% for Google and 18% for Walmart.
Muse has more than 10 million users in the US and Canada since launching in September, Reuters reported. Amazon blocked Muse that month and last year sued Perplexity to prevent similar access, a case Reuters says is still ongoing. Jassy said on the third-quarter call that “few customers are doing much shopping this way yet”, as we reported from that call. Meta has since added phone calls to Muse for paying subscribers.
By building the most extensive store of product data, Amazon could ensure that agents shopping for users are compelled to seek products on Amazon.com rather than other retailers’ sites, the people told Reuters. Employees describe the plan as an expansion of Amazon’s Buy for Me and Shop Direct services, which list products sold on other sites. For the basics, see agentic commerce explained.
The Structural Read
As we read the Reuters report next to Forter’s data, Amazon is building on both sides of the agent question. It blocked Muse, the agent that on Forter’s data bought at 74% of stores with any agentic order, and its document plans product data so complete that agents would need to come to Amazon. Morgan Stanley’s figures, as Reuters gives them, show the gap the plan targets: 57% of shoppers start research on Google against 16% on Amazon, about 3.6 times as many by our arithmetic.
Amazon internal document, as reported by Reuters
“Many Rufus answers are factually wrong or incomplete, but customers lack expertise to detect these errors”
Three Implications
FOR BRANDS The document says Coca-Cola and Bayer, among others, have signed up to contribute product data, and Amazon would also crawl other retailers’ websites, Reuters reported.
FOR AI SHOPPING AGENTS Amazon blocked Muse in September and is still in court with Perplexity over similar access, Reuters reported; the people said fuller data could compel agents to seek products on Amazon.com.
FOR GOOGLE On the Morgan Stanley figures Reuters cited, 57% of online shoppers start product research on Google and 16% on Amazon; the document names Google as one of the places customers leave Amazon to research.
The framework argues: “The transformer is a commodity now; what changes between paradigms is where intelligence accumulates and who can capture it.”
As we read it, World Product Graph is a bet on where shopping intelligence accumulates: in the product data an agent needs, held by Amazon. The framework predates the Reuters report and does not discuss Amazon’s plan.
Share of stores on Forter’s network with any agentic order where each agent bought, 1 August to 7 October 2026, from our piece on Forter’s data. Amazon blocked Muse in September, Reuters reported. The 3.1 times ratio is our arithmetic.
What Is Not Established
We did not see the document; every figure from it comes through Reuters. Amazon says the usage figures are significantly outdated, and the current Rufus search volume is not in the reporting we read. Reuters could not determine whether the 1 October date is realistic, and the people said World Product Graph could be modified or aborted. The link between the plan and blocking agents is the people’s account, as Reuters reported it.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
An internal Amazon document reviewed by Reuters plans World Product Graph, data on virtually every item for sale, and puts Rufus at about 83 million searches a week, a figure Amazon says is significantly outdated. The people familiar told Reuters the plan is also a defence against AI shopping agents such as Muse, which Amazon blocked in September.
95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. We read Reuters’ 13 November exclusive in full through its MarketScreener syndication; reuters.com itself served a bot check. We did not see Amazon’s internal document; every figure and quote from it comes through Reuters, as does Amazon’s statement. The Forter figures come from our 13 November piece on Forter’s agent order data. The 3.6 times and 3.1 times ratios are our arithmetic. We did not contact Amazon, Meta or the brands named.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
