The World Wide Web Consortium this week announced a joint workshop with GS1 on the future of e-commerce in a world where AI agents increasingly sit between web content and the end users who consume it. The event, titled "E-commerce for Humans and AI Agents," is scheduled for September 8 and 9, 2026, in Zurich, Switzerland, with remote participation available. Google is hosting the event.

The announcement, published June 1, 2026 on the W3C website, opens a formal call for participation with a submission deadline of June 26, 2026.

What the workshop is about

According to W3C, web content has historically been designed for humans, even when pages also expose structured data through APIs or embedded JSON-LD using vocabularies such as schema.org and the GS1 Web Vocabulary. That model is under pressure. Large language models and AI agents are becoming a new intermediary between that content and end users - summarizing search results, following links, and potentially supporting users in online activity before a person ever visits a page directly.

The core question the workshop sets out to examine is what this shift means for the people and organizations that create, curate, and publish web content. According to W3C, the event aims to share experiences of creating content with AI agents in mind, focusing specifically on e-commerce. What practices are most effective? What are the pitfalls? What adjustments do content creators need to make to maximize the return on investment in time, energy, and skill?

The impetus is commercial. E-commerce is the particular focus, though the organizers expect the discussion to draw on broader perspectives related to content creation and publication generally.

The technical backdrop: MCP, UCP, and ACP

Three protocols sit at the center of the technical conversation the workshop will address: MCP (Model Context Protocol), UCP (Universal Communication Protocol), and ACP (Agent Communication Protocol). According to W3C, the workshop asks what practices help retailers, brands, content creators, curators, LLM and AI agent providers, SEO practitioners, and online data providers remain visible and useful through these protocols, while keeping the web truly open rather than forcing content creators into closed or semi-closed ecosystems.

MCP has become a dominant technical pattern for connecting AI agents to external data sources and platforms. Anthropic introduced MCP as an open standard in November 2024, and adoption across the marketing technology sector accelerated sharply throughout 2025 and into 2026. Amazon opened its advertising APIs through MCP in February 2026, and Google proposed WebMCP at I/O 2026 in May as a browser-level open web standard that allows developers to expose structured tools - JavaScript functions and HTML forms - so browser-based AI agents can execute complex tasks without relying on screenshot parsing and DOM manipulation.

Chrome 149 subsequently opened a WebMCP origin trial, a first concrete schedule attached to what had previously been a proposed standard. That trial is relevant context for the September workshop: the infrastructure questions W3C and GS1 want to examine are not hypothetical. Several of the protocols under discussion are already in active development or limited deployment, and the commercial stakes for retailers and brands are rising.

The W3C's interest in this space is not new. In February 2026, it held a three-day virtual Workshop on Smart Voice Agents, which identified eight unresolved standards gaps covering interoperability, privacy, hallucination control, and accessibility. The official report, published on March 31, 2026, mapped a roadmap of work that W3C Community Groups, upcoming events, and a possible new W3C activity could take forward. The September e-commerce workshop extends that pattern of using convened events to map an emerging standards territory.

The JSON-LD and schema.org context

The structured data layer the workshop references - particularly JSON-LD using schema.org vocabulary and the GS1 Web Vocabulary - has existed for years. Schema.org was launched in 2011 by Google, Bing, Yahoo, and later Yandex to create a common set of schemas for structured data markup on web pages. More than 10 million websites use schema.org markup. The GS1 Web Vocabulary, maintained by the same GS1 organization co-hosting this workshop, extends schema.org specifically for product-level commercial data.

The challenge the workshop addresses is that this infrastructure was built for a different discovery environment. Search engines crawled pages, read structured data, and presented results to human users who then clicked through. AI agents change that sequence. They may read structured data and never send a user to the page at all, or they may act on product information without the kind of referral traffic that structured data was originally designed to generate.

Google's Gary Illyes highlighted related infrastructure fragility in analysis published by PPC Land in June 2026, noting that the web's reliance on JavaScript-heavy pages creates significant problems for AI agents attempting to extract structured information. Server-side rendering, semantic HTML, and clean document structure - technical SEO fundamentals - have taken on new significance as the systems consuming web content multiply and diversify.

For e-commerce specifically, the implications are direct. PPC Land reported in May 2026 that Google flagged eight specific Merchant Center feed attributes required for visibility in AI Mode and AI Overviews, with precise character limits and documented formatting errors that break data ingestion. That kind of attribute-level specification is exactly the territory the W3C and GS1 workshop wants to examine systematically.

Who can participate

According to W3C, attendance is free for invited participants and is open to people who can contribute relevant experience, use cases, technical perspectives, implementation experience, research findings, or standardization questions. The organizers aim for diverse participation across five specific communities: content creators and curators of all kinds; LLM and AI agent providers, including in-browser AI agents; brands making products to sell, and retailers selling them; search engine optimization practitioners; and online data providers.

Two participation pathways exist. A Position Statement is appropriate for people who want to suggest a topic, share a perspective, or be considered for a presentation, panel, demonstration, or breakout discussion. An Expression of Interest is appropriate for people who want to attend and contribute to discussion but do not have a specific topic to propose.

According to W3C, a good position statement should be a few paragraphs long. It should include the submitter's background in the main topic areas, what they would like to contribute or discuss, how their contribution relates to e-commerce for humans and AI agents, relevant use cases or standardization gaps, and links to related supporting resources if available.

Submissions can be made through GitHub - using either the Position Statement template or the Expression of Interest template - or by email to group-ecommerce-agents-pc@w3.org. For email submissions, the subject line should follow the format "[Position Statement] - Your Name / Organization" or "[Expression of Interest] - Your Name / Organization." Submitters should indicate whether they prefer to attend in person or remotely, to help organizers plan venue capacity and remote participation support.

The deadline for both types of submission is June 26, 2026. Acceptance notifications will go out on August 1, 2026. An initial agenda will be published on August 17, 2026.

The Programme Committee

The workshop's Programme Committee includes representatives from a notable range of organizations. Phil Archer from GS1 serves as co-chair, alongside a second chair listed as TBD. Committee members represent a broad cross-section of the commercial and technical web: Mark van Berkel from Schema App, Trish Burgess from Airbnb, Jean-Luc Di Manno from Fime, Wei Ding from Huawei, Anssi Kostiainen from Intel, Alex Jansen and Michael Linck from Google, Natalia Macia from Bosch, Gavin Shenker from Visa, Andrew Smith and Riley Strong from OpenAI, Vasilii Trofimchuk from Block Inc., Andrea Volpini from WordLift, and Léonie Watson from Tetralogical. A further committee member is listed as TBD.

The presence of OpenAI representatives alongside Google personnel on the same programme committee is notable. Both organizations are developing AI systems that interact with web content at scale, and both have commercial interests in how the standards that govern that interaction evolve. Visa and Block Inc. add a payments-layer perspective to discussions that will inevitably touch on checkout flows and transaction data. Airbnb and Bosch represent large-scale product and inventory publishers with structured data operations at the commercial end of the market.

Why this matters for marketing and advertising

The workshop sits at the intersection of several forces that have been reshaping digital advertising and search marketing throughout 2025 and 2026. The shift toward AI-mediated discovery - where a user asks a question and an agent assembles an answer from multiple sources rather than presenting a list of links - changes the visibility economics that digital marketing has operated on for two decades.

Skepticism about the commercial viability of AI shopping agents has been documented since at least October 2025, with analysts identifying structural challenges including retailer incentives against AI intermediation and high e-commerce return rates. Amazon and Shopify - which collectively control more than 50% of the US e-commerce market - have reportedly blocked AI agents to maintain discovery ownership and protect retail media businesses. The W3C workshop does not resolve these commercial tensions, but it provides a venue where technical standards work can proceed in a way that at least acknowledges them.

The standards that emerge from processes like this one have long time horizons. The W3C updated its own Process Document in August 2025 to streamline how standards advance through the development pipeline - removing the Proposed Recommendation phase and introducing Charter Refinement - but even a simplified standards process operates on a timescale of years. The September 2026 workshop is positioned as a scoping and alignment event, identifying shared questions and standardization opportunities rather than producing finished specifications.

For SEO practitioners and e-commerce operators, the practical implication is that the infrastructure decisions made in 2026 - how product data is structured, which vocabularies are used, how pages expose their functionality to agents - will shape visibility in agent-mediated discovery for years to come. The workshop offers a channel to influence how those standards are defined.

Questions about the workshop can be directed to the Programme Committee at group-ecommerce-agents-pc@w3.org.

Timeline

  • June 2, 2011 - Schema.org launched by Google, Bing, Yahoo, and Yandex as a collaborative structured data vocabulary initiative
  • November 2024 - Anthropic introduces Model Context Protocol (MCP) as an open standard for connecting AI agents to external data sources
  • February 25-27, 2026 - W3C holds its Workshop on Smart Voice Agents, identifying eight unresolved standards gaps in interoperability, privacy, hallucination control, and accessibility; PPC Land coverage
  • February 2026 - Amazon opens its advertising APIs through MCP in open beta; PPC Land coverage
  • February 15, 2026 - Chrome's WebMCP discussed as a developing standard for structured browser-based agent interaction; PPC Land coverage
  • March 31, 2026 - W3C publishes official report from its Smart Voice Agents workshop, mapping eight unresolved cross-cutting issues and next steps; PPC Land coverage
  • May 15, 2026 - Google publishes a guide for website owners on optimizing content for generative AI features in Search; PPC Land coverage
  • May 19, 2026 - Google proposes WebMCP at I/O 2026 as a proposed open web standard for exposing structured tools to browser-based AI agents; PPC Land coverage
  • May 19-25, 2026 - Chrome 149 origin trial opens for WebMCP, the first concrete deployment schedule for the standard; PPC Land coverage
  • June 1, 2026 - W3C publishes announcement of the joint W3C/GS1 Workshop on E-commerce for Humans and AI Agents, scheduled for September 8-9, 2026, in Zurich
  • June 3, 2026 - Gary Illyes analysis on JavaScript complexity as an AI agent obstacle published; PPC Land coverage
  • June 26, 2026 - Deadline for submitting Position Statements and Expressions of Interest to the workshop Programme Committee
  • August 1, 2026 - Acceptance notifications sent to submitters
  • August 17, 2026 - Initial workshop agenda published
  • September 8-9, 2026 - Hybrid workshop in Zurich, hosted by Google

Summary

Who: The World Wide Web Consortium (W3C) and GS1, with Google as host, and a Programme Committee including representatives from OpenAI, Airbnb, Visa, Bosch, Intel, Huawei, Tetralogical, Block Inc., WordLift, Schema App, and Fime.

What: A two-day hybrid workshop titled "E-commerce for Humans and AI Agents," examining how retailers, brands, content creators, SEO practitioners, and online data providers can remain visible and useful as AI agents become intermediaries between web content and end users, with a focus on protocols including MCP, UCP, and ACP, and structured data vocabularies including schema.org and the GS1 Web Vocabulary.

When: The workshop runs September 8-9, 2026. The call for participation is open now, with a submission deadline of June 26, 2026. Acceptance notifications are expected by August 1, 2026, and an initial agenda will be published August 17, 2026.

Where: Zurich, Switzerland, with remote participation available. Hosted by Google.

Why: AI agents are increasingly mediating e-commerce discovery, interacting with product data and web content in ways that existing structured data infrastructure - built for human-facing search - was not designed to support. W3C and GS1 are convening the workshop to identify shared questions and standardization opportunities, with the goal of ensuring the web remains open and that content creators retain visibility and commercial value through the transition to agent-mediated discovery.