The cost of being anonymous: E-E-A-T in AI search

We changed nothing about our articles except who signed them, and the content score moved eighteen points. Same sentences, same structure, same evidence. The only difference was a name, a role, and a link to a real profile.

That is the cheapest lesson in this article, so it goes first. What follows is why it works.

The four letters, without the SEO fog

E-E-A-T is Experience, Expertise, Authoritativeness, Trustworthiness. In practice they answer four different questions, and most sites only try to answer one.

Experience — have you actually done this? Not studied it. Done it. First-hand contact with the thing you are describing. This is the newest addition to the framework and the hardest to manufacture, because it shows up in specifics: numbers you could only have if you ran the process, failures you would not invent, details nobody quotes from a competitor's blog.

Expertise — do you know the subject? Demonstrated by accuracy and depth, undermined by vagueness. A page that hedges every claim reads as someone who has read about the topic rather than worked in it.

Authoritativeness — does anyone else say you know it? This one is not on your website. It lives in other people's citations, mentions and links.

Trustworthiness — is there a reason to believe you? Contact details that resolve, claims that are checkable, and someone accountable for what is written.

Why anonymity costs more in AI search than in classic search

Classic search ranks a list. A user who lands on an anonymous page can evaluate it themselves — look at the design, check the About page, decide.

An AI answer removes that step. The system quotes a claim and stands behind it. If the claim turns out to be wrong, the platform absorbs the reputational damage, not you. So the calculation shifts: given two pages with equivalent content, the one with an accountable author is a lower-risk source to repeat.

You are not being judged on writing quality. You are being judged on whether someone can be pointed at.

The minimum viable setup

This is genuinely small. Four things:

1. A real name on every article, visible in the byline, not buried in a footer.
2. A Person node in your structured data, with a stable @id, linked as the article's author and as the organisation's founder or employee.
3. A link from that person to an external profile that confirms they exist and do this work.
4. A short, honest bio — real background, no invented credentials. See below.

That is a couple of hours of work and it is the highest-return trust change available to a small site, because everything else in the trust layer depends on other people.

Do not invent the biography

The temptation with the first E is to inflate it. Resist, for two reasons.

The practical one: invented credentials are checkable. An AI system that cross-references your claimed background against external profiles and finds nothing has learned something about your reliability — and it is not what you hoped.

The strategic one: an honest unusual background is more persuasive than a generic impressive one. "Ten years in digital marketing" is unfalsifiable and forgettable. A specific, verifiable path — even a sideways one — gives a reason to trust the method rather than the title.

What to do this week

Open any article on your site. If it is signed by your company rather than a person, you are leaving the cheapest trust signal on the table.

Add the name. Add the Person markup. Link it to a profile that exists. Then move on to the parts that take months — the ones that depend on other people mentioning you.