E-E-A-T for AI: How to signal expertise so models cite you

In 2023, Google formalised its E-E-A-T framework - Experience, Expertise, Authoritativeness, Trustworthiness - as the evaluative lens its quality raters use to assess content. It turns out this framework is an almost perfect map of what AI models look for when deciding whether to cite a source.

The intuition is the same in both cases: a human or a model reading a page is not just evaluating the content, it is evaluating the source of the content. Who is making this claim? Do they have a track record? Can I find them if this turns out to be wrong? AI models trained on human-generated text have absorbed the same trust heuristics humans use.

The 5 on-page E-E-A-T signals AI models look for

1. Author markup

The most impactful single change for most sites. Add a visible author name to content pages - not just a byline for aesthetics, but also in schema markup. The combination of a visible author, a link to an author page, and schema:author in the page's structured data gives AI models three independent signals that a real, nameable person is accountable for the content.

{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "Your article title",
  "author": {
    "@type": "Person",
    "name": "Jane Smith",
    "url": "https://yoursite.com/about/jane-smith",
    "sameAs": ["https://linkedin.com/in/janesmith"]
  },
  "datePublished": "2026-04-20",
  "dateModified": "2026-04-20"
}

2. An About page with specific credentials

An About page linked from the main navigation tells AI crawlers that this is an established organisation, not an anonymous content farm. The specifics matter more than the length: years of operation, team size, named founders, specific domain expertise, and verifiable credentials (certifications, publications, awards) are the signals AI models look for.

Generic phrases like 'We are passionate about helping businesses grow' carry no weight. Specific claims like 'Founded in 2019, we have run AEO audits on over 50,000 websites' are citable facts that AI models can use to frame a citation.

3. Organization schema with address

Adding a physical address (or at minimum a city and country) to your Organization schema is a strong trustworthiness signal. It ties the content to a real-world entity that can be held accountable. This is especially important for B2B SaaS, finance, health, and legal content - categories where AI models apply higher trust thresholds before citing.

4. External citations in your content

Linking out to reputable primary sources - academic papers, official documentation, government data, established news organisations - signals that your content is part of a web of knowledge, not a self-contained assertion. AI models are more likely to cite a page that cites others because it signals the author's claims are verifiable.

The inverse is even more valuable: becoming a primary source other sites cite. Publishing original research - a proprietary statistic, survey, or benchmark - is the strongest trustworthiness signal available, because it cannot be found anywhere else for an AI model to cross-check.

5. Freshness and date signals

Publish dates and last-modified dates in your schema (datePublished, dateModified) and in your HTTP headers matter for queries about current topics. An undated page looks like it could be from any era; a page with explicit dates gives the AI confidence it is not citing stale information.

The address element: an overlooked AEO signal

The HTML <address> element is semantically meaningful - it marks contact information for the nearest article or the whole page. Including it in your footer or About page, alongside Organisation schema with the same address, provides two consistent, machine-readable signals of real-world existence. Very few sites use <address> correctly, which makes doing so a meaningful differentiator.

E-E-A-T for YMYL content

If your site covers Your Money or Your Life topics - finance, health, legal, safety - the E-E-A-T bar is higher. AI models apply a stricter trust threshold before citing sources that could affect a user's wellbeing. For these sites, professional credentials (licensed financial advisor, board-certified physician, licensed attorney) should appear explicitly in author schema and visible text, and content should be reviewed and dated.