New Machine-Readable Layers Emerge for AI Agent Interaction

Image credit: Search Engine Journal

Global web infrastructure is developing new machine-readable layers, including Google‘s Open Knowledge Format (OKF) and Agentic Resource Discovery (ARD) specifications, to facilitate interactions between artificial intelligence agents and content discovery.

These emerging standards represent a significant shift toward a parallel web infrastructure designed for AI, potentially altering how search engine optimization (SEO) operates and how AI systems process information, industry observers said.

Google introduced OKF and ARD, which add distinct layers to the web’s existing structure, Search Engine Journal reported. These specifications aim to make websites more accessible and understandable for AI.

The publication identified six layers for preparing a site for AI: Crawlable HTML, Schema.org/Structured Data, LLMs.txt, MCP/WebMCP, Open Knowledge Format (OKF), and Agentic Resource Discovery (ARD).

OKF is described as a collection of markdown files containing YAML frontmatter. Its primary purpose is to enable internal knowledge sharing among AI agents, rather than serving as a replacement for traditional web crawling or a public content retrieval system, according to Google’s John Mueller.

Agentic Resource Discovery (ARD) is an open specification detailing how AI agents can locate and verify tools, skills, and other agents across the internet. This specification focuses on defining capabilities rather than content, Francois Vanderseypen explained.

Suganthan Mohanadasan, an SEO expert, suggested that new formats like OKF and LLMs.txt might experience an adoption trajectory similar to Schema.org. He noted that early adoption could provide the most value before AI systems become more advanced and explicit markup becomes less critical or shifts to a verification role.

The development of these new machine-readable layers indicates a strategic move by Google and other stakeholders to build a web environment optimized for the capabilities and needs of AI agents.

Source: Search Engine Journal