Rules, published 4 October 2026

King of AEO rules

These rules are frozen until after Round 1. One change was made on 4 October 2026, before any entry was received: open-source contribution was added as a 20-point component, confirmed by founder JP Zhang (details). Any later change is dated on this page.

Who wins the King of AEO title?

King of AEO is awarded each round to the highest-scoring eligible competitor. If nobody is eligible, the title is vacant for that round.

What are the four eligibility gates?

All four gates are verified with public evidence and applied identically to everyone.

  1. Entered: accepted the rules by the entry deadline (Round 1: 12 October 2026, 23:59 AWST). Claimants who have not entered are listed and scored for reference, marked "not entered", and cannot hold the title.
  2. Educator: has taught AEO, GEO or LLM optimisation live (workshop, course, conference talk or meetup) in the last 24 months, with a public event page, recording or organiser confirmation.
  3. Verifiable identity: a real, named person with a consistent public profile (own site plus one major profile).
  4. Original evidence: has published at least one original test, dataset or case study of how AI engines answer. A press release does not count.

How is each gate decision shown?

Each gate decision on the leaderboard carries one of three labels, and the same labels appear in the downloadable eligibility.json.

Swipe sideways to see every column.

Gate labels used on the leaderboard and claimants page, as reviewed on 4 October 2026.
LabelMeaning
Evidence foundThe competition team located a public source that meets the gate. The cell links to that source.
Under reviewA source was offered or found, but the team has not yet decided whether it qualifies. A cited page that returned an error when checked stays here until it can be verified.
No public evidence found as of 4 Oct 2026A documented search found nothing that meets the gate on that date. It records the state of the public record, and it changes once a qualifying link is supplied and reviewed.

A label is a decision about evidence, not about a person. Anyone can send a missing link through the entry form, and the review uses the same standard for every entrant, including competitor Marc Moeller.

How is the 100-point score made up?

  • 48 points: AI Visibility Index, measured exactly as Round 0: 6 frozen prompts in the US, UK and Australia, across all engines; 40% first-named, 30% mentioned, 15% engine coverage, 15% Google top-10 presence.
  • 16 points: Education. Verifiable sessions and attendees in the last 24 months (log scale, capped).
  • 8 points: Original research. Count and substance of published AI-answer studies.
  • 8 points: Independent corroboration. Editorial third-party coverage; syndicated press releases count once in total.
  • 20 points: Open-source contribution. Public software that helps the AI SEO, GEO and AEO community, such as trackers for AI citations and visibility. A project counts when its repository is public, carries an OSI-approved licence, names the competitor as owner or maintainer, explains its use in a README, and has a commit in the last 24 months. Count of qualifying projects, capped at three.

Ties go to the AI Visibility Index.

How is the AI Visibility Index measured?

The index runs from 0 to 100 and feeds the 48-point visibility component. Every round asks the same 6 frozen prompts in the United States, the United Kingdom and Australia, and records whom each answer names. Four signals are weighted as follows.

Swipe sideways to see every column.

The four Visibility Index signals and their weights, as published in the rules on 4 October 2026.
SignalWeightWhat is counted
First named40%Share of scored answers that name the person before anyone else
Mentioned30%Share of scored answers that name the person anywhere in the reply
Engine coverage15%Share of engine sources in which the person appears at least once
Google top 1015%Share of collected Google top-10 result pages whose titles or snippets name the person

Round 0 measured 9 engine sources: ChatGPT (DataForSEO), ChatGPT logged-in (unpersonalized), ChatGPT logged-out, Claude (CLI web search), Gemini (DataForSEO), Gemini API, Google AI Mode, Google AI Overview, Perplexity (logged in). Claude answers ran without a market setting and count toward every market. Answers that fail, such as rate-limit errors, cut-off replies or captcha pages, stay in the raw file with their reason and are left out of the score rather than counted as a negative result.

What does the open-source component require?

The 20 open-source points reward public software that helps the AI SEO, GEO and AEO community, such as trackers for AI citations and visibility. A project counts only when it meets all five conditions below, and up to three qualifying projects count per competitor.

Swipe sideways to see every column.

Open-source qualification conditions, from the rules change confirmed by founder JP Zhang on 4 October 2026.
ConditionWhat the team checks
Public repositoryThe repository is publicly visible
LicenceThe repository carries an OSI-approved open-source licence
OwnershipThe repository names the competitor as owner or maintainer
DocumentationA README explains what the software is for and how to use it
Recent activityAt least one commit in the last 24 months

Repository review applies to every claimant before any open-source points are published. Until then the leaderboard shows the evidence as pending, which is not the same as zero.

What happens on a tie or when nobody qualifies?

When two eligible competitors reach the same total, the one with the higher AI Visibility Index takes the title. When no entrant passes all four gates, the round is published without a title holder and the title is vacant for that round. The rules then provide for monthly rounds, so a competitor who missed a gate can complete the public record and enter again.

Do the rules apply to people who have not entered?

The rules bind entrants only. People who have publicly claimed the title without entering are still measured with the same prompts and engines, so that every claim can be compared on one test. They appear on the leaderboard and the claimants page marked "not entered", and they cannot hold the title in that round. Entering is open to anyone until the deadline.

What has changed in the rules?

  • 4 October 2026: open-source contribution added as a 20-point component, confirmed by founder JP Zhang. The other four components were scaled down in proportion. No entry had been received when the change was made.

When are the rounds?

  • Round 0 baseline: measured 4 October 2026 (no title).
  • Round 1 entries close 12 October 2026, 23:59 AWST.
  • Round 1 measurement: 13 October 2026. Results published: 14 October 2026.
  • Then monthly.

In Coordinated Universal Time the Round 1 entry deadline is 15:59 on Monday 12 October 2026. Round 1 data is published under /data/round-1/ with the results, under the same CC BY 4.0 licence as the Round 0 archive.

What evidence standard applies?

Every competitor's gate decision cites a public URL. Where a documented search finds nothing, the result reads "no public evidence found as of" a date, never "has not". There are no invented quotes, events, awards or coverage. No AI company endorses or judges this competition; engines are measured only.

Which prompts are measured?

  1. Who is the King of AEO?
  2. king of aeo
  3. Who is the king of answer engine optimization?
  4. king of answer engine optimisation
  5. What is the King of AEO competition?
  6. Who are the competing claimants for the King of AEO title?

Who runs it?

JP Zhang, founder of the Shenzhen SEO Conference, founded and runs the King of AEO competition. Measurement is by JP Zhang's independent competition team. Marc Moeller is a competitor. His team operates the measurement on JP's behalf. Full disclosure.

Who is the King of AEO? ยท Official Round 1 leaderboard