I installed EntityMap on this site.

EntityMap is a new open standard for declaring a structured, entity-first index of website knowledge to AI systems. JSON file at the root of the domain (/entitymap.json), HTML companion (/entitymap.html), v1.0 spec released 27 May 2026. It treats the sitemap.xml analogy seriously. Sitemap.xml says “these URLs exist.” EntityMap says “these entities exist, here is how they relate, here is where the evidence lives.” I shipped both files plus a footer link this morning. Mine is at /entitymap/ and /entitymap.json.

The reason I took it seriously this early is the names attached. Dixon Jones is one of two editors. He founded Majestic. He runs InLinks. He has done entity-based search work for over a decade. Putting his name on a brand-new spec earns it a slot on the “worth tracking” list. Fred Laurent is his Waikay co-founder. The spec is published under CC BY 4.0, which keeps it open even though Waikay is a commercial product that will eventually monetise the tooling around it.

It’s been a day since this is introduced. So it means zero AI systems read EntityMap as of writing. (ChatGPT, Claude, Perplexity, Google AI). The GitHub repo has 11 stars. The Waikay reference generator is “Coming Soon” on Waikay’s own site. This is Wave 0 of adoption. Shipping the file does not get me cited anywhere new today. It does not move AI Overview impressions. It does not show up in Search Console.

What I am betting is small. Two hours of work to write a well-formed entitymap.json declaring 9 entities (me, my two companies, and the topics I want to be associated with: AI Agent SEO, Schema Markup for LLMs, MCP, llms.txt, Markdown for Agents, Search Journey Optimisation). The HTML companion takes another hour. Footer link is one line. Total cost is a half-day for the optionality of being early if EntityMap catches on.

This matches the framing in The Three Lives of Schema Markup and the agent-ready piece. Schema took a decade to earn out. Sitemap.xml took years to get respected at scale. Most protocol-layer work is registration not advertising. If you treat EntityMap as a citation lever you will be disappointed. If you treat it as one more piece of entity infrastructure that may compound quietly in the background, the maths is fine.

Two practical notes if you want to try it yourself. Write the JSON manually or use a non-Waikay generator. Anchoring your entity declarations to a single vendor’s product workflow is the kind of small thing you regret in two years. And be selective. EntityMap is not a content index. Declare the entities you want to be known for, not every blog post you have ever written. Tight is better than complete.

For the agent-ready post I will leave it as is for now. EntityMap will get folded into the protocol stack list once a real consumer (any major LLM, Chrome Lighthouse audit, Cloudflare default) acknowledges support. Until then it sits in this note, in my footer, and in my /entitymap.html page as quiet infrastructure.

Watching, not betting.