The Mutual Flatness Theorem

Two Views. One Page.

Left: how you see this page. Right: how an AI agent reads it -- a semantic graph extracted from the JSON-LD @graph embedded in the HTML. Same content. Radically different perception.

Human View
GEO Research March 2026

What is Generative Engine Optimization?

Generative Engine Optimization (GEO) is the practice of optimizing web content so that AI-powered search engines -- like ChatGPT, Perplexity, Gemini, and Claude -- cite your content in their generated answers.

Unlike traditional SEO, which focuses on ranking in a list of 10 blue links, GEO focuses on being cited as a source in AI-generated responses. The shift is fundamental: from ranking to citation.

The Key Factors

01

Structured Data (@graph)

JSON-LD with @graph pattern and Wikidata entity linking. This is what you see on the right -- the machine-readable layer of this page.

02

Citation Signals

Brand mentions now correlate 3x more with AI citations than backlinks (0.664 vs 0.218). Authority is built differently in the AI era.

03

Machine-Readable Architecture

robots.txt, llms.txt, sitemap.xml -- the files that tell crawlers exactly what to read and how to read it. GEO starts here.

04

Knowledge Graph Presence

Wikidata, Wikipedia, branded entity pages. AI engines need to resolve your brand as a known entity before they can cite you.

The Experiment

Crawl Atlas is a live GEO experiment. Every page on this site is built with GEO-first architecture: structured data, entity linking, AI-friendly content structure. We track which AI engines crawl us, how they parse our content, and whether they cite us in their answers.

The graph on the right is real. It's generated from the actual JSON-LD embedded in this page's HTML. What you see as readable paragraphs, the machine sees as a network of typed entities and relationships.

> TIP

Rotate the graph on the right by dragging. Hover over nodes to see entity details. Zoom with scroll. Each glowing node is a schema.org entity, each line is a relationship.

Machine View JSON-LD @graph -- Live