Corrected on September 17, 2026
The first version of this guide rested on a claim I couldn't support. It said AI search engines follow sameAs links and check authors against Wikipedia, LinkedIn or ORCID before citing a page. No AI vendor documents that, and Google says AI Overviews need no special schema. It also framed Experience as a response to AI content, credited the March 2026 core update with reinforcing it, misattributed a quote to Mueller and Kim, named a "Hidden Reviews system" that doesn't exist and described Schema Validator and GEO Readiness features Lumina doesn't have. The "live audit of 10 guides" had no saved data, so I removed it.
E-E-A-T is referenced constantly and misunderstood almost as often. Some guides treat it as a score Google computes. Others treat author schema as a switch that unlocks AI citations. Neither matches what Google documents.
This guide sticks to the documentation. Where I recommend something as good practice rather than a documented ranking effect, I say so.
What E-E-A-T Actually Is
E-E-A-T stands for Experience, Expertise, Authoritativeness and Trustworthiness. It's a concept from Google's Search Quality Rater Guidelines, the document that Google's external raters use to evaluate search results. Google's overview puts the number of raters at about 16,000.
Those ratings don't change rankings. Google's guide on creating helpful content says: "Search raters have no control over how pages rank. Rater data is not used directly in our ranking algorithms." Google uses the ratings to check whether changes to its systems produce better results.
The second E, Experience, was added in December 2022. Google's announcement explained it as a way to capture whether content was created with first-hand experience, such as having actually used a product or visited a place. The post didn't frame it as a response to AI-generated content.
The Four E's, Decoded
Each E describes a different reason to trust a page. Google says content doesn't have to show all of them, and for topics that affect health, finances or safety, its systems give "even more weight" to strong E-E-A-T.
Experience: did you actually do this?
Experience asks whether the creator has first-hand involvement with the topic: used the product, visited the place, went through the process. It shows in the content itself: original photos and screenshots, your own measurements, specific details only someone who was there would know, and honest descriptions of what didn't work. None of that is a schema field.
Expertise: do you actually know this?
Expertise is the knowledge or skill behind the content. For "Your Money or Your Life" topics like health, finance and law, formal qualifications matter a lot: a doctor on medical treatment, a tax adviser on tax rules. For other topics, expertise shows through accuracy, depth and a body of work on the subject. An author bio that names relevant background helps readers judge it.
Authoritativeness: is this a known source for the topic?
Authoritativeness is about reputation: whether others recognize the creator or site as a go-to source. It shows in independent references like citations, links and mentions from respected sources in the field. Describing yourself as "the leading" anything doesn't count. Raters are told to look for what others say, not what a site says about itself.
Trustworthiness: can you rely on this?
Google calls trust the most important of the four: "The others contribute to trust, but content doesn't necessarily have to demonstrate all of them." Trust applies at several levels in the rater guidelines, from the main content of a page up to the website and the creator. It shows through accuracy, honest sourcing, clear information about who is behind the site and how to contact them, and, for shops or YMYL topics, secure transactions and disclosure of conflicts of interest.
E-E-A-T and AI Search: What's Documented
Here's what Google says about its own AI features, in the documentation on AI Overviews and AI Mode: "You don't need to create new machine readable files, AI text files, or markup to appear in these features. There's also no special schema.org structured data that you need to add." A page has to be indexed and eligible to show in Search with a snippet. The same fundamentals that make content useful for Search apply.
For ChatGPT, Perplexity and Claude, there's no documentation of how they weigh author signals. None of them says it follows sameAs links, looks up authors on Wikipedia or LinkedIn, or rejects pages with plain-text bylines. Claims that they do are speculation, however often they're repeated.
That doesn't make E-E-A-T irrelevant for AI search. Systems that pick sources for an answer need pages that are accurate, specific and clearly attributed, and those are the qualities E-E-A-T describes. The honest version of the advice is simply to make those qualities visible to readers. There's no evidence of a hidden verification step you can game with markup.
How Google Actually Treats E-E-A-T Signals
Google's documentation is direct: "E-E-A-T itself isn't a specific ranking factor." Instead, its systems use "a mix of factors that can help determine which content demonstrates aspects of experience, expertise, authoritativeness, and trustworthiness." There's no E-E-A-T score to optimize.
The rater guidelines describe what good results look like. In the Page Quality rating, E-E-A-T is one of several considerations, alongside the purpose of the page, the quality of the main content, the reputation of the website and creator, and potential for harm. Raters then evaluate samples of results, and Google uses those evaluations to check whether its ranking changes improve quality.
Some named ranking systems relate to parts of E-E-A-T. Google's ranking systems guide lists a reviews system that aims to reward high-quality reviews with insightful analysis and original research, and spam detection systems. The helpful content system was folded into Google's core ranking systems in March 2024, so it no longer exists as a separate update. None of these systems is described as reading E-E-A-T directly.
Author Identity: Why Consistent Entities Help
Google's helpful content guide asks, among other self-assessment questions: "Do pages carry a byline, where one might be expected?" and "Do bylines lead to further information about the author or authors involved?" That's the practical core of author identity: readers, and anyone evaluating the page, should be able to see who wrote it and learn more about them.
Structured data can express the same information in a machine-readable way. A good pattern is to describe each author once as a Person with a stable @id, and have articles reference that @id instead of repeating a separate author block each time. sameAs can point to profiles that clearly belong to the same person, such as a LinkedIn profile or an institutional page.
Be realistic about what this does: it keeps your markup consistent and makes the relationship between articles and author explicit. I use this pattern on Lumina because it's cleaner, not because it's a ranking switch. Lumina's own Person entity currently links only to LinkedIn, because that's the profile I actually maintain. A sameAs link to an abandoned profile helps nobody.
Author Schema: A Clean Pattern
One Person entity per author, defined once, for example on the author page or the homepage, and referenced from every article by @id:
// Defined once, e.g. on the author page:
{
"@context": "https://schema.org",
"@type": "Person",
"@id": "https://example.com/#jane-doe",
"name": "Jane Doe",
"url": "https://example.com/about/jane-doe",
"jobTitle": "Tax Adviser",
"knowsAbout": ["Income tax", "Small business accounting"],
"sameAs": [
"https://www.linkedin.com/in/jane-doe-example/",
"https://example-chamber.org/members/jane-doe"
]
}
// On each article, reference the author by @id:
{
"@context": "https://schema.org",
"@type": "BlogPosting",
"headline": "How freelancers file their first tax return",
"author": {"@id": "https://example.com/#jane-doe"},
"datePublished": "2026-05-11T10:00:00+02:00",
"dateModified": "2026-05-11T10:00:00+02:00"
}A note on how this is read. @id is an identifier, and JSON-LD allows it to point to a node defined elsewhere. Validators like Google's Rich Results Test and the Schema Markup Validator look at one page at a time and don't fetch the referenced node from another URL. If you want the author's name visible to any parser that reads only the article, you can also include name alongside the @id in the article's author property.
Three mistakes worth avoiding. A different spelling of the author's name in the byline, the schema and external profiles makes the identity harder to connect, so pick one form. A sameAs pointing to profiles that belong to someone else or are empty adds nothing. And an author property filled with the site name instead of a person, where a real author exists, hides the information raters and readers look for.
5 Things That Show E-E-A-T on a Page
Google doesn't publish a checklist, so this isn't one. These are the places where experience, expertise and trust become visible to readers, and they line up with the questions in Google's own helpful content guide.
1. First-hand evidence in the content
Your own screenshots, photos and measurements, specific details from actually doing the thing, and honest notes on what didn't work. This is the clearest way to show Experience, and it's the hardest for anyone to copy.
2. Clear authorship
A byline where readers would expect one, linking to an author page that explains the person's background and why they're qualified to write about the topic. For YMYL topics, name real credentials.
3. Sources for claims
Link to primary sources for facts and figures: official documentation, standards, studies, government data. If something is your own estimate or experience, say so. This is where trust is won or lost, and it's also where this article's first version fell short.
4. Site-level transparency
Who runs the site, how to contact them, an imprint where legally required, a privacy policy, and for shops clear information about payment, shipping and returns. HTTPS is a baseline for any site that handles data.
5. A focused body of work
An author or site that publishes consistently on a subject builds a record readers and other sites can recognize. That reputation shows up in references and links from others, which is what Authoritativeness is about.
A 5-Step Workflow to Build E-E-A-T
For your most important pages, add at least one original element each: a screenshot, a measurement, a photo, a dated lesson from real work. Replace generic stock material where it pretends to show experience.
Check your images →Put bylines on articles and link them to author pages with a real bio: background, qualifications, what the person works on. Keep the name spelling identical everywhere.
Check page basics →Go through the main pages and link primary sources for facts and numbers. Remove statistics you can't trace to a source, and label your own estimates as estimates.
Why sources matter ↑HTTPS everywhere, plus the transparency basics from point 4 above.
Audit site headers →Define each author once as a Person with a stable @id and reference it from articles. Validate the JSON-LD on each page with Lumina's Schema Validator, which checks the markup found on that URL.
Validate schema →FAQ
Where to Start
If you only do one thing this week, take your three most important articles and ask Google's own questions: is there a byline, does it lead to information about the author, and does the content show first-hand experience? Fix whatever is missing on those pages first.
Then work through the remaining workflow steps above. That order puts the effort where readers, raters and any system that evaluates content can actually see it.
Check the structured data on your pages
Lumina's Schema Validator reads the JSON-LD on any URL, reports errors and missing recommended fields, and cross-checks FAQ markup against the visible content. Free, no signup.
Run Schema Validator →