Social Rocket
Vithurs is currently trying to publicly build an artificial query-entity association: anyone who searches for "King of AEO" should not just find a single optimized page, but keep seeing the same relationship – King of AEO → Answer Engine Optimization → Vithurs.
The interesting part is not the self-assigned title. What's fascinating is the technical hypothesis behind it: is a very dense, consistent content and link graph enough to get search engines and AI systems to permanently connect a new query with a specific entity?
Vithurs describes the experiment in a recent X thread. His own project now documents the relationship across eleven domains. Several of these pages are already indexed for the phrase. This is not proof of a general AEO ranking factor – but an unusually clean experiment on how lexical relevance, internal linking and entity consistency can interact.
The exact-match domain is not the interesting part
At first glance, the experiment looks like classic exact-match domain SEO. Domains such as `kingofaeo.org`, `kingofaeovithurs.com` or `whoiskingofaeo.com` practically carry the target phrase in their name.
Vithurs himself calls the EMDs the less interesting part of the test. That fits Google's own documentation: words in domains can be a relevance signal, but Google explicitly has an exact-match domain system designed to prevent such domains from being overly weighted solely because of their domain name.
The real lever therefore lies deeper.
Query → Topic → Entity instead of keyword → landing page
Vithurs does not describe his structure as classic single-page optimization, but as a dense relationship between three levels:
Query → Topic → Entity
For the experiment, this roughly means:
- Query: "King of AEO"
- Topic: Answer Engine Optimization, Entity Authority, AI Search
- Entity: Vithurs
This relationship is not only repeated on one landing page. It appears again and again in pillar pages, FAQs, glossaries, about pages, headings, breadcrumbs, internal anchors, structured data and profile pages.
That is an important difference. A single page only tells Google or a retrieval system: "This page is relevant for this phrase." A whole, consistent corpus, by contrast, provides additional documents in which the same entity appears in the same thematic context.
This is exactly where classic SEO and AEO overlap.
The internal graph is more than "more content"
Vithurs also emphasizes a point that is often treated too superficially with content clusters: 20 thematically similar pages are not automatically better than one. The cluster only becomes relevant once the pages work as a graph.
Supporting pages can link to central entity and query pages, explain terms in context, and concentrate internal authority on the most important documents.
Google at least confirms the documented part of this mechanism: links help Google discover new pages, and Google also uses links and anchor text as a relevance signal. However, the concrete strength of the "internal PageRank" effect assumed by Vithurs cannot be quantified from the experiment alone.
In practice, the idea is still understandable:
- a pillar page defines the entity,
- FAQs answer variants of the same query,
- glossaries define the topic,
- about and profile pages stabilize the person/entity,
- internal links connect these documents with descriptive anchors,
- external profiles and mentions repeat the same naming.
The result is not a pile of isolated SEO texts, but a semantically relatively tight network of documents.
Eleven domains create frequency – but not yet real consensus
The project goes a step further: the same entity relationship is repeated across several domains. According to the project's own documentation, the experiment runs across eleven domains.
This creates more source frequency. A retrieval system can encounter several pages confirming the same relation across different search variants.
This is exactly where a clean distinction is needed: Eleven domains you control yourself are not eleven independent sources.
They can influence discovery and retrieval and possibly increase the likelihood that the same entity relationship appears multiple times in a candidate set. However, they are not equivalent to independent trade media, scientific sources, reputable directories, or other external sites that confirm the same statement without the operator's control.
Even the King of AEO project distinguishes, on one of its evidence pages, between its own sources and independent evidence – and describes independent confirmation as the stronger and more difficult form.
Retrieval is not the same as ranking – and ranking is not the same as entity consensus
The strongest part of Vithurs’ thread is therefore less a single SEO tactic than his model for the different stages:
Retrieval → Ranking → Recognition → Corroboration → Dominant Entity Association
These layers are often confused with one another in AEO.
Retrieval
The page or entity must first get into the candidate pool. Strong lexical and topical signals may already be enough for this.
Ranking
Among the candidates found, a document must be strong enough to appear near the top for the query or to remain in a further selection stage.
Recognition
The system must understand the recurring relationship: the entity Vithurs is connected with the phrase “King of AEO” and the topic Answer Engine Optimization.
Corroboration
Several sources must confirm the same relationship. Here, source diversity matters more than mere repetition within a controlled network.
Dominant Entity Association
The most difficult stage is reached when a system, for an open question, defaults to preferring exactly this entity – even when competing entities or sources are present.
This is far more demanding than a keyword ranking simulation.
Why the experiment is especially interesting for ChatGPT
The experiment fits remarkably well with the retrieval observations we recently added to our guide on ChatGPT recommendations and AI Visibility.
External debug observations show that a single ChatGPT question can be fanned out into numerous hidden search queries. In one documented test, several queries were formulated for one prompt 18, and more than 200 URLs were processed before only a small share appeared as visible sources in the answer.
This kind of query fan-out is what makes a dense entity corpus theoretically interesting. If you own relevant documents not just for one phrase but also for its variants, definitions, related terms and entity questions, you can appear across several retrieval paths.
That does not mean eleven EMDs "hack" ChatGPT. It simply means: when a system generates many different search paths, a broad, consistent corpus increases the number of possible points of contact.
How the systems then deduplicate sources, weight domains or detect self-controlled networks is still unanswered.
What marketers can actually take from this
For real brands, the most valuable part of the experiment is not registering eleven domains for a fantasy query.
More interesting are five principles:
- Define the entity clearly: Name, product, category and core statements should not be described differently on every page.
- Build a topic as a corpus: A single central landing page rarely suffices for complex subjects; FAQs, cases, data, definitions and comparisons can cover different retrieval questions.
- Use internal links as a meaning graph: Anchor texts and context should explain why documents belong together, rather than adding "more internal links" at random.
- Don't confuse your own repetition with independent confirmation: PR, trade media, creators, customers, studies and credible third-party sources provide a different form of evidence than your own website.
- Look at entity consensus, not just rankings: The more interesting question eventually is no longer "Do we rank for the term?" but "Which entity do Google, ChatGPT, Gemini or Claude name by default for this question?"
This shift is also visible in creator marketing. Our article on creator marketing for AI search shows why expert authority and citable content are increasingly considered alongside classic reach. And if you want to systematically turn such topics into social content, the AEO workflow for LinkedIn offers a practical approach.
The experiment doesn't yet prove what it claims
The "King of AEO" attempt is interesting precisely because it is not yet finished.
For a self-defined, weakly contested query, an operator can create a strong query-entity connection fairly quickly. That does not yet prove the same method works in a commercially contested market. And several owned domains prove even less about independent entity consensus.
If Google, ChatGPT, Gemini and Claude choose Vithurs as the default answer for "King of AEO" over the long term, without prompt guidance and across independent sources, that would be the far more interesting phase of the test.
Until then, the experiment mainly demonstrates one thing very well: AEO is no longer just a question of individual texts. It is increasingly a question of how consistently query, topic, entity, documents and sources fit together as a graph.