Link
deadline:
Call for participation
“Web content” has historically meant “designed for humans”: HTML pages with combinations of text, imagery, audio, video and more. All styled with CSS, made accessible with the judicious use of ARIA, and made interactive with JavaScript. “Data on the Web” typically refers to some sort of data exchange between clients and servers through APIs. Of course the distinction between the two can be very blurred.
Infinite scrolling Web pages make repeated calls to APIs and the data is then rendered for human consumption; billions of HTML pages have embedded structured data, usually as a block of JSON-LD and using the schema.org vocabulary.
Browser APIs, CSS and JavaScript functionality mean that rich Web content is readily available. But the working assumption is that the end result will be consumed by people.
Ultimately, that remains the case, however, there is a new intermediary between the content and the end user: LLMs and AI agents. Already, this is having a dramatic effect on browsing habits. In a typical search engine user experience nowadays, the list of links returned by a search query is still presented to the user, but those links come below an AI-generated summary that for many users is as far as they get.
While these links may no longer be followed by the user, they were followed by the bots that gathered input data to the LLM’s training set, or are followed in real time by AI agents supporting the user in their online activity.
This presents whole new sets of challenges to content creators who want to cater to this new usage paradigm and find new audiences through protocols - generic ones like Model Context Protocol (MCP), or ecommerce specific ones like the Universal Commerce Protocol (UCP) from Google et al, and Open AI’s Agentic Commerce Protocol (ACP).
The old challenge of “how can an online retailer ensure that potential customers are aware of my offer” becomes “what do I have to do to be noticed by AI shopping protocols?”
The old challenge of “how can my carefully curated and referenced content” be found among the noise becomes “how can I help AI to direct readers to my carefully curated and referenced content?”
But one question remains unchanged: how can the Web remain truly open so that content creators are not forced to choose a closed or semi-closed ecosystem?
Goals
This workshop aims to share experience of creating content with AI Agents in mind. What are the practices that are most effective? What are the pitfalls? What shifts do content creators need to make in order to maximize the return on the investment in time, energy and skill?
The impetus for the workshop is from ecommerce which will be a particular focus of the event. However, the discussion is expected to be informed by broader input related to content creation and publication.
Intended audience
The following list is indicative only.
- Content creators and curators of all kinds
- LLM and AI Agents providers, including in-browser AI Agents
- Brands making products to sell, and retailers selling them
- Search Engine Optimization practitioners
- Online data providers
Participation
Participation details are still to be confirmed.