Amazon's World Product Graph Plan Targets AI Shopping Agents and E-Commerce Rivals

Amazon Moves to Control Global Product Data as AI Shopping Agents Threaten Traditional E-Commerce Platforms

San Francisco, November 14, 2026: Amazon is exploring an ambitious initiative to build a comprehensive database of products sold across the global retail market, including merchandise that is not available directly on its marketplace, as artificial intelligence-powered shopping agents threaten to reshape online commerce.

The proposed initiative, internally known as World Product Graph, would combine product information supplied by manufacturers with data gathered from other retailers' websites.

According to internal planning documents reported on November 13, the project is designed to help shoppers research, compare and select products without leaving Amazon's platform.

The initiative reflects a growing competitive challenge from AI shopping assistants capable of independently searching multiple websites, comparing prices and completing purchasing tasks on behalf of consumers.

Such services could weaken the traditional e-commerce model, in which customers visit retail websites, browse product listings and interact with advertisements before making purchases.

Amazon is responding on two fronts: exploring ways to expand its own product information infrastructure while restricting certain external AI agents from accessing its marketplace.

However, World Product Graph remains a proposed initiative rather than a confirmed product launch. Amazon has described the underlying planning document as outdated and cautioned that it does not necessarily reflect the company's current plans.

Amazon's World Product Graph: Key Developments

What Is Amazon's World Product Graph?

World Product Graph is a proposed system intended to organise information about products available throughout the global retail market.

Amazon already maintains a substantial product catalogue covering merchandise sold through its own marketplace.

However, consumers frequently leave Amazon to research product specifications, compare models, examine manufacturer websites or read independent reviews.

The proposed system would seek to reduce that need.

It could provide information about products that Amazon does not directly stock, allowing customers to compare alternatives across a broader selection.

For example, a shopper researching a laptop could potentially identify its latest version, compare specifications and evaluate competing models within Amazon's interface.

The project would represent an expansion from operating an online marketplace toward becoming a comprehensive product discovery platform.

Its commercial significance lies in the possibility of influencing purchasing decisions before customers select a retailer.

Amazon Wants to Become a Central Destination for Product Research

Traditional e-commerce platforms compete not only to complete transactions but also to attract customers at the beginning of their shopping journey.

Consumers frequently research products through search engines, manufacturer websites, online reviews and social platforms.

Amazon's proposed database is intended to bring more of that research activity into its own ecosystem.

A 2025 Morgan Stanley research report cited in the November 13 reporting found that 57% of online shoppers began product research on Google, compared with 16% on Amazon.

However, Amazon held a stronger position when shoppers already knew which product they wanted to purchase.

In those circumstances, 41% began searching on Amazon, compared with 23% on Google and 18% on Walmart.

The figures illustrate the distinction between product discovery and purchase intent.

World Product Graph would aim to strengthen Amazon's position in the earlier research stage, where Google remains a major competitor.

How Amazon Plans to Build Its Product Database

The proposed system would combine several sources of product information.

Manufacturers could supply structured details about their products, including specifications, versions and relationships between models.

Amazon could also gather publicly available information from other retail websites.

The objective would be to create a more complete understanding of products and their alternatives.

According to the reported planning document, Coca-Cola and Bayer were identified as companies contributing product information.

The document also described discussions involving a potential trial with Procter & Gamble.

However, these references come from internal planning materials, and the extent of current participation has not been independently established.

The proposal raises important questions about how Amazon would obtain, update and use product information belonging to other businesses.

Product Data Could Become a Major Competitive Asset

Product information is increasingly important to digital commerce.

A product listing can include specifications, dimensions, compatibility information, photographs, pricing and availability.

For conventional online shopping, this information helps customers evaluate individual items.

For AI shopping systems, structured product data serves an additional purpose.

It enables software to compare alternatives and make recommendations based on a customer's requirements.

An AI assistant searching for a washing machine, for example, may need to understand capacity, energy consumption, dimensions and installation requirements.

Incomplete or inaccurate data can produce unsuitable recommendations.

A comprehensive product database could therefore improve shopping assistance while reducing the cost of processing complex queries.

For Amazon, control over a larger product information network could become an important competitive advantage.

AI Shopping Agents Challenge the Traditional E-Commerce Model

AI shopping agents are software systems that can perform purchasing-related tasks on behalf of consumers.

Unlike conventional search engines, which primarily return links and product information, advanced agents can navigate websites and carry out multiple steps.

A customer might ask an agent to identify a suitable product within a budget, compare retailers and prepare a purchase.

Depending on the system and permissions, the agent may also complete checkout.

This changes how retailers interact with customers.

Instead of attracting shoppers directly to their websites, merchants may increasingly compete to be selected by automated assistants.

The shift could reduce the importance of traditional website navigation and some forms of digital advertising.

For large marketplaces, maintaining access to customers during product discovery becomes strategically important.

Amazon Restricts Access for Competing AI Agents

Amazon has taken steps to restrict certain external AI shopping agents from accessing its retail platform.

The company blocked Meta's Muse shopping agent from its marketplace in September 2026.

It has also been involved in litigation concerning Perplexity's AI-powered shopping capabilities.

Amazon argues that automated systems operating on its platform must respect its access requirements and protect customer information.

The company has raised concerns about agents that interact with its website without clearly identifying themselves.

These restrictions illustrate a broader conflict between established retailers and emerging AI shopping services.

Retailers want to control how their platforms are accessed, while AI developers seek to help users shop across multiple websites.

The outcome could influence the technical and commercial rules governing agent-driven commerce.

Amazon and Perplexity Dispute Access to Shopping Data

Amazon's dispute with Perplexity has become an important example of the challenges surrounding autonomous shopping.

Perplexity's Comet browser includes an assistant capable of navigating websites and performing tasks at a user's direction.

Amazon has objected to the assistant's access to its marketplace without the authorisation and identification it requires.

The company has argued that undisclosed automated activity can create security, privacy and customer experience concerns.

Perplexity's position centres on the ability of users to direct software to perform tasks on their behalf.

The dispute raises questions about the boundaries between authorised user activity and automated third-party access.

Those questions extend beyond Amazon because similar agents could interact with banks, travel platforms and other digital services.

The litigation remains an important development for the future of agentic commerce.

Meta's Muse Adds Competitive Pressure

Meta's Muse shopping agent is another example of growing competition in automated commerce.

According to the November 13 reporting, the service attracted more than 10 million users across the United States and Canada after its September launch.

Its growth illustrates consumer interest in AI-powered shopping assistance.

For Amazon, the concern is not simply that competing agents can recommend products.

It is that they may direct purchasing activity toward retailers offering better prices, more complete information or alternative products.

This could weaken the relationship between Amazon and customers who previously began shopping directly on its platform.

Amazon's decision to block Muse reflects its determination to retain control over access to its marketplace.

However, consumer adoption of independent shopping agents remains an evolving market.

Amazon's Own AI Shopping Assistant Becomes Central to Strategy

Amazon has developed its own AI shopping capabilities through Rufus, which it began referring to as Alexa for Shopping in May 2026.

The assistant helps customers research products, compare alternatives and receive recommendations.

It is designed to make shopping more conversational.

Instead of searching through multiple product pages, customers can describe what they need and ask follow-up questions.

However, the usefulness of these systems depends heavily on the quality of their underlying information.

Amazon's internal planning document acknowledged problems involving incomplete or inaccurate product answers.

The proposed World Product Graph could help address these limitations by expanding the information available to the assistant.

Amazon Says Alexa for Shopping Is Approximately 97% Accurate

The reported internal document raised concerns about the accuracy of Amazon's existing shopping assistant.

It also referred to substantial query volumes and the expense of operating AI-powered search.

Amazon disputed the relevance of those figures.

The company said the planning document contained outdated usage information and stated that Alexa for Shopping was approximately 97% accurate globally.

Amazon CEO Andy Jassy also said the assistant had been used by more than 400 million people during the preceding year.

These statements indicate the scale of Amazon's investment in AI-assisted shopping.

However, the company has not publicly established that World Product Graph will be launched in its proposed form.

The accuracy figure is also Amazon's own reported measurement, rather than an independently established industry benchmark.

Better Product Data Could Reduce AI Operating Costs

AI shopping systems can be expensive to operate because they must interpret customer requests and process product information.

Complex questions may require multiple searches, comparisons and reasoning steps.

A more structured product database could simplify these tasks.

Instead of repeatedly gathering information from different websites, a shopping assistant could retrieve organised product relationships from an established system.

This may reduce computational requirements for some queries.

It could also improve consistency when comparing specifications or identifying newer product models.

However, maintaining a comprehensive database would involve substantial costs.

Amazon would need to collect, validate and regularly update product information across many categories and retailers.

The overall economic benefit would depend on whether these efficiencies exceed the cost of maintaining the system.

Amazon's Advertising Business Faces a New Challenge

Advertising is an important part of Amazon's commercial model.

Brands and merchants pay to promote products within Amazon's shopping environment.

These advertisements help businesses reach consumers who are actively considering purchases.

AI shopping agents could change that process.

If customers rely on independent assistants to select products, they may spend less time browsing sponsored listings.

An agent might compare products based on price, specifications and customer preferences without reproducing the advertising experience of a conventional marketplace.

This creates uncertainty about how retail advertising will evolve.

Amazon has indicated that the impact of external shopping agents on its advertising business remains limited.

Nevertheless, the emergence of agent-driven purchasing presents a potential long-term challenge to traditional sponsored product discovery.

Amazon CEO Andy Jassy Downplays Immediate Threat

During Amazon's third-quarter earnings discussion, CEO Andy Jassy indicated that relatively few consumers were completing purchases through independent AI shopping agents.

He also said the effect on Amazon's advertising operations had been minimal.

The comments suggest that the company does not currently view external agents as a major source of lost business.

However, the reported World Product Graph proposal indicates that Amazon is evaluating how shopping behaviour could change.

Technology platforms often invest in emerging capabilities before those capabilities become widely adopted.

The distinction between current commercial impact and future strategic risk is important.

AI agents may still represent a relatively small share of online purchasing, but their capabilities are developing rapidly.

Google and Walmart Face a Broader Product Discovery Battle

Amazon's proposed initiative could intensify competition with Google and Walmart.

Google has a strong position in product research through its search engine and shopping services.

Walmart operates a major retail and e-commerce business with its own customer relationships and product catalogue.

Amazon competes with both companies at different stages of the purchasing process.

World Product Graph would seek to strengthen Amazon's role before a customer decides where to buy.

If successful, it could reduce the need to visit competing search engines or retail websites for product information.

However, rivals are also investing in AI-powered search, recommendations and shopping experiences.

The competitive outcome will depend on product accuracy, consumer trust, retailer participation and purchasing convenience.

Buy for Me and Shop Direct Offer a Foundation

Amazon has already introduced services intended to connect customers with products sold outside its own marketplace.

Buy for Me and Shop Direct allow shoppers to discover merchandise available through other retailers.

These initiatives expand Amazon's role beyond products stocked or sold directly through its platform.

World Product Graph could build on that approach by providing more comprehensive information about external merchandise.

A customer might research an item through Amazon even when the final purchase involves another retailer.

This could allow Amazon to participate in more stages of the shopping journey.

However, the proposed database and existing purchasing services are separate initiatives.

Their future integration has not been confirmed.

Product Data Access Raises Competition Questions

The emerging dispute over shopping data raises questions about competition and platform access.

Amazon wants to protect its marketplace from certain external automated agents while exploring ways to gather information about products sold elsewhere.

These activities involve different technical and legal arrangements.

Nevertheless, they highlight the increasing importance of controlling product information.

Retailers may argue that restrictions are necessary to protect customer privacy, platform security and commercial systems.

AI developers may argue that automated shopping tools should be able to act on behalf of consumers.

The balance between these interests is likely to become an important issue for digital commerce.

Regulatory treatment may differ across jurisdictions and depend on the circumstances of each case.

Retailers May Need to Adapt to Agent-Driven Shopping

The growth of AI shopping agents could change how retailers present products online.

Traditional e-commerce strategies emphasise website design, search rankings, advertising and customer engagement.

Agent-driven commerce may place greater importance on accurate product information and machine-readable catalogues.

Retailers could need to ensure that product specifications, pricing and availability are accessible through authorised technical systems.

Brands may also need to reconsider how they communicate product advantages when recommendations are generated by software.

However, human shopping behaviour will remain important.

Consumers may continue visiting retailers directly for browsing, customer service, brand experiences and purchasing decisions.

The transition is therefore likely to involve a combination of conventional and AI-assisted commerce.

What Amazon's Strategy Means for Brands and Sellers

For manufacturers and marketplace sellers, a comprehensive product information system could create opportunities and challenges.

More complete product data may improve discoverability and reduce confusion between similar models.

Brands could benefit when AI systems accurately explain product features and compatibility.

However, greater dependence on a single platform for product discovery could also increase concerns about commercial leverage.

Sellers may have less influence over how products are compared or recommended.

The importance of accurate catalogue information could increase.

Businesses may need to maintain consistent product records across marketplaces, brand websites and other sales channels.

The commercial impact will depend on how Amazon develops and governs any future system.

World Product Graph Launch Remains Uncertain

The internal planning document contained an aspirational launch date of October 1, 2027.

However, that date should not be interpreted as a confirmed release schedule.

Amazon said the document was outdated and did not necessarily reflect its current development plans.

The company also emphasised that it regularly explores new technologies and ways to improve its product catalogue.

Internal planning documents can describe proposals that are subsequently revised or abandoned.

World Product Graph could therefore change substantially before any public launch.

It may also never become a commercial product in the form described.

The distinction matters because the reported initiative reflects strategic exploration rather than an announced operational rollout.

Outlook for AI Shopping and Global E-Commerce

The development of AI shopping agents is creating a new competitive environment for digital retailers.

Product information, recommendation accuracy and customer access are becoming increasingly important.

Amazon's proposed World Product Graph illustrates how established marketplaces may respond by expanding their data infrastructure.

At the same time, restrictions on external AI agents show that access to retail platforms is becoming contested.

Google, Walmart, Meta, Perplexity and other technology companies are pursuing different approaches to shopping assistance.

The future market may involve competition between retailer-controlled assistants and independent agents operating across multiple platforms.

The outcome will depend on consumer adoption, commercial partnerships, technical standards and legal decisions concerning automated access.

For Amazon, the central challenge is maintaining its position as a leading destination for shopping as the process of finding and purchasing products changes.

Conclusion

Amazon is exploring a major expansion of its product information infrastructure as AI shopping agents challenge traditional e-commerce platforms.

The proposed World Product Graph would seek to organise information about products sold across the global retail market, including items not available directly on Amazon.

The initiative could improve product comparisons, strengthen Alexa for Shopping and reduce dependence on external websites for product research.

It could also help Amazon defend its advertising and marketplace businesses against independent shopping agents.

The company's restrictions on certain external agents, including Meta's Muse, and its legal dispute with Perplexity demonstrate the growing commercial importance of controlling access to shopping platforms.

However, World Product Graph remains an unconfirmed proposal, and Amazon has cautioned that the reported internal document does not necessarily represent its current plans.

The broader development signals a significant change in e-commerce competition.

As AI systems increasingly assist consumers with product discovery and purchasing, the businesses controlling reliable product information may gain greater influence over how online shopping evolves.