E-E-A-T for AI Models: Author Identification Is Key

What E-E-A-T means — in brief

Google originally introduced E-E-A-T for its human quality raters: external reviewers who assess search results against quality criteria and whose feedback feeds into algorithm updates. The four dimensions are:

• Experience: Does the author have personal, hands-on experience with the topic? "I built this myself" beats "I read about it".
• Expertise: Does the author possess deep subject knowledge? Credentials, training, demonstrable specialization.
• Authoritativeness: Is the author recognized as an authority by others? External references, citations, "known for".
• Trustworthiness: Is the website transparent? Clear contact information, no misleading behavior, data protection.

The problem with SEO for a long time was this: these signals were optimized primarily for human reviewers. Pretty About pages, impressive bio texts, certificates as JPGs. What got overlooked: AI models read this differently.

How AI models assess authority

Unlike human quality raters, an AI does not read your About page "with understanding" — it processes tokens. That means: everything you want to communicate has to be encoded in a machine-readable way. A beautifully designed biography helps little if the underlying metadata is missing.

Concretely: when Claude or Perplexity handle a query and have to decide which source is trustworthy, they analyze primarily three machine-readable signals:

Signal 1: Schema.org Person markup

JSON-LD with @type: "Person" gives an AI structured information about the author — name, job title, employer, areas of expertise (knowsAbout), external profiles (sameAs). It is the direct, machine-readable way to say: "This person is an expert in X."

Signal 2: Consistent authorship across URLs

An AI does not assess just a single page — it assesses the domain. When Thomas Leonhardt consistently appears as the author on the About page, in every blog post header, in the footer, and in the Article schema, a coherent authority signal emerges. Inconsistency — sometimes "T. Leonhardt", sometimes "the author", sometimes no name at all — fragments the signal.

Signal 3: Linking to verifiable profiles

The sameAs property in Schema.org lets you connect the author with external, verified profiles: LinkedIn, GitHub, XING, a personal website. AI models can use these links to verify authorship and gauge expertise.

Schema.org Person — a complete example

The following JSON-LD snippet belongs on the /en/about/ page. It is the author's machine-readable identity card — fully filled in, with no field left out:

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Person",
  "name": "Thomas Leonhardt",
  "url": "https://forgeproject.eu/en/about/",
  "jobTitle": "Founder & AI Orchestrator",
  "description": "Thomas Leonhardt founded FORGE in 2020 and has been building autonomous AI agent systems ever since. 40+ completed projects, specializing in GEO and multi-agent orchestration.",
  "worksFor": {
    "@type": "Organization",
    "name": "FORGE",
    "url": "https://forgeproject.eu"
  },
  "knowsAbout": [
    "Generative Engine Optimization",
    "AI agent orchestration",
    "Multi-agent systems",
    "GEO strategy",
    "Schema.org markup"
  ],
  "sameAs": [
    "https://www.linkedin.com/in/thomas-leonhardt-forge/",
    "https://forgeproject.eu"
  ],
  "image": "https://forgeproject.eu/og-image.png",
  "alumniOf": {
    "@type": "Organization",
    "name": "FORGE"
  },
  "nationality": {
    "@type": "Country",
    "name": "Germany"
  }
}
</script>

{
  "@type": "Article",
  "author": {
    "@type": "Person",
    "name": "Thomas Leonhardt",
    "url": "https://forgeproject.eu/en/about/"
  }
}

<a href="/en/about/" rel="author">Thomas Leonhardt</a>
<span>Founder & AI Orchestrator, FORGE</span>