Agentic commerce requires two key components to operate accurately and autonomously: product identity and interoperable protocols.
-
What Is Agentic Commerce?
Agentic commerce is a shopping experience where AI agents act on a shopper’s intent to discover, compare, and ultimately purchase products. AI agents, or software that can interact with webpages and applications to perform actions based on a user’s input, are entering a commerce ecosystem built on a foundation of product identity and interoperability. For AI agents to function properly at scale, they must tap into that foundation
On the surface, agentic commerce is a simple automation. A shopper inputs their intent into their AI assistant, and it purchases the product. For example, they ask their agent to order a pair of sneakers from a particular brand, with parameters around size, style, and price. The agent autonomously executes this task, and in moments the shopper receives receipt of purchase and a notification of impending shipment. In a recent survey from Adobe, 39% of shoppers say they have used AI for online shopping, with 85% of them saying it improved their experience. While simple on the surface, behind the scenes, the process is a careful orchestration of shopper intent, product identification, and interoperable systems that allow AI agents to navigate and act.
-
Related Articles
How Does Agentic Commerce Work?
Product Identity
Agents need to understand the user’s intent and link that to real-world products that they can then discover, compare, and purchase. The Global Trade Item Number® (GTIN®) is a globally unique identifier that provides a source of truth for a product’s identity and can serve as a reference point as AI agents navigate the web in search of the answer to their user’s query. When expressed as a GS1 Digital Link, a GTIN can be resolved by AI agents to authoritative product information published by the brand owner or other trusted sources. Keyword searches, images, and product attributes are other ways an AI agent can determine the right product to purchase based on their input criteria.
How can a business create GTINs for its products?
A GTIN provides a globally unique identity for a product. Businesses can license individual GTINs or use a GS1 Company Prefix to create and manage GTINs across a growing product portfolio. Both options help establish the trusted product identity that supports commerce across channels.
Interoperable Protocols
Once the agent identifies the product, it needs to navigate multiple websites to discover and compare products across various sellers. For this reason, AI companies have developed protocols that enable interoperability. Google’s Universal Commerce Protocol (UCP), Stripe and OpenAI’s Agentic Commerce Protocol (ACP), Google’s Agent Payments Protocol (AP2), and Anthropic’s Model Context Protocol (MCP) are all different protocols that govern aspects of how an agent transacts across websites
Product identity and interoperable protocols are the consistent threads across agentic commerce implementations. Data standards are the foundation supply chains already run on; they are the common language that lets product and inventory information move cleanly across trading partners through a combination of manual and automated processes. Standards provide the visibility and governance that AI agents need throughout their journey.
What Is the Agentic AI Journey?
The Source: Where Product Identity Begins
A brand determines a product’s identity before an AI agent interacts with it. Nevertheless, it is a key piece of the process that can influence the accuracy and efficiency of agentic commerce downstream. Accurate product data is essential at the origin point, as the information compounds at every step that follows. Each downstream system inherits the brand’s first answer, right or wrong. Therefore, grounding product identity in a GTIN reduces the risk of misidentification later in the supply chain
A key difference between traditional shopper search and agentic commerce is that the AI agent returns a single answer. There are no longer hundreds of results for the shopper to scroll through and still find a product on page two. In agentic commerce, the agent removes a product with bad identification data from consideration entirely.
Product information management (PIM) systems, warehouse management systems (WMS), and enterprise resource planning (ERP) systems are databases where product identity may already live. A brand needs to reconcile the data into one authoritative identity that gives agentic commerce a source of truth.
The Translation: How Product Identity Moves
From its origins at a manufacturer through the distributor to the retailer’s shelf, every product move is a translation, and standards are what keep those hand-offs accurate. A Global Location Number (GLN) identifies each party and place in the chain, making it clear who is sending what and where it is going. Yet each time a partner transfers product data, something can be lost in translation. The Global Data Synchronization Network® (GDSN®) exists to close that gap by keeping product data in sync across trading partners instead of leaving each one to maintain its own copy
GDSN is one way data moves between trading partners. What matters just as much is that the attributes mean the same thing on both ends of the hand-off. The GS1 Global Data Model (GDM) defines a common set of product attributes, including descriptions, features and benefits, weights and dimensions, and other key product details. That common structure helps ensure product data is represented consistently across trading partner systems, whether a company syncs through GDSN, sends a spreadsheet, or publishes to a portal. A shared classification like GS1's Global Product Classification (GPC) adds another layer of consistency by providing a common way to group products. Together, GDM and GPC give AI agents both a consistent vocabulary for describing a product and a consistent way to categorize it.
The Discovery: How an Agent Finds and Compares
Once a brand establishes the product identity, the agent still has to find it. This is where interoperable protocols enter the journey. An agent can rarely complete a shopper’s request through a single site, so it moves across multiple sellers, catalogs, and marketplaces to discover options and compare them against the shopper's intent. Without interoperability, the agent’s journey stops at boundaries between systems, and the agent cannot complete its task
Protocols move the agent between systems, but the agent still must read what it finds when it arrives. Much of the product information on the open web is written for human shoppers, so an agent parsing a page of marketing copy must infer product identifiers, attributes, and offers from the page rather than read them directly as structured data. The GS1 Web Vocabulary addresses this by expressing GS1 Standards as linked data on the web, extending schema.org with commerce-specific product properties. When a brand publishes product data using the GS1 Web Vocabulary, an agent can read the identifier, the attributes, and the offer as structured facts. This is the practical end of the GS1 Digital Link mentioned earlier. The GS1 Digital Link takes an agent to authoritative product information, and the Web Vocabulary provides the language that allows the agent to interpret it consistently.
Protocols such as UCP, ACP, AP2, and MCP are the connective tissue that lets an agent carry a request from one system to the next without starting over each time. Identity tells the agent what the product is, the vocabulary tells it how to read what it finds, and protocols tell it how to move. Without all three parts working together, agentic commerce struggles with identification and accuracy and requires human intervention. With all three parts in place, the agent moves the right request from one seller to the next, weighing options against the shopper's intent until it settles on the one to buy.
The Purchase: How a Shopper Buys the Product
When it is time for the shopper to make the purchase, the AI agent has made its choice. Whether that choice is right depends on the product identity underneath it. For brands that do not have the underlying product identity, this is where sales are lost, and they are lost quietly.
When the brand ensures that the product has a unique, standardized identifier like a GTIN, the agent can match the shopper’s intent with the right product, the right brand, and the right size. The transaction clears, the confirmation arrives, and the shopper never has to think about the underlying data that made it possible.
An agent that cannot reconcile a product's identity and attributes across sellers must pivot to address the shopper’s intent. The agent picks a substitute product elsewhere, and the shopper never learns what they missed. There is no abandoned cart to analyze, no bounce to measure, no second impression to win the customer back. The price and availability of the product may be competitive, but without the proper identity data linked to it, it is far more likely to disappear in a crowded marketplace. A shopper who once compared options on a results page now delegates that judgment to an agent, and the agent rewards the product it can identify with confidence.
Conclusion
In agentic commerce, products are not just ranked; they are selected. Trusted unique product identity helps determine whether they are selected at all. When a shopper hands their intent to an AI agent, every step that follows, from discovery and comparison through purchase and even return, runs on the product identity and attributes the brand established at the source. However, if product identification is poor or if related data about the product is incomplete or inconsistent, the agent may disregard the product entirely. Product identity and interoperable protocols are the consistent threads that carry a request cleanly from one system to the next, and standardized data is what makes that identity trustworthy at every hand-off.
Technology platforms and AI companies are building the protocols. Brands control the product data those agents rely on. The most valuable move a brand can make now is to ensure the product identity that feeds AI is clean, consistent, and grounded in a GTIN before an agent ever goes looking for it. This is a critical juncture where there is still time for brands to bring the conversation back to their teams and ask a simple question: When an AI agent interacts with our brand’s product data today, what is it reading, and will that make its way to the shopper?
Interested in Preparing for Agentic Commerce?
Whether you're looking for more information, want to exchange ideas with peer thought leaders, or are interested in exploring how GS1 Standards can make your product identity and data agent-ready, we want to hear from you. Agentic commerce is moving quickly, and the brands that ground their product data in global standards now will be the ones agents can find, trust, and choose later.
