Do named authors get cited more by AI? What the evidence shows

Nobody has published a controlled study showing that putting a named author on a page makes ChatGPT, Perplexity, or Google's AI Overviews cite it more often. What does exist is explicit guidance from Google that pages should carry a byline where readers expect one, leading to real information about the author, and that trust is the most important part of how quality is judged. So the honest answer is: bylines are a well-founded trust practice with an unproven direct effect on AI citation. They are still worth doing, for reasons that hold whether or not that effect is ever measured.

A number we took down

An earlier version of this article, published on the previous version of our site, said that pages with a named author were cited 4.1 times more often than identical pages without one, and that adding bylines lifted the median AI Overview citation rate in health, finance, and legal pages from 4 percent to 16 percent. When we prepared this rewrite we went looking for the methodology behind those figures and could not find one we were willing to publish. There was no record of the page sample, how "identical pages" were matched, which engines were queried, or how many runs each result was based on.

A statistic you cannot reproduce is a claim, not a finding, so we removed both numbers. We say this plainly because the AI search field is full of confident multipliers with no method behind them, and our own editorial standard is that every quantitative claim links to a primary source. When we have enough opted-in audit data to measure authorship effects properly, we will publish the method and the raw counts together, including if the effect turns out to be small.

What Google actually says about authors

Google's guidance on helpful content asks three questions about any page: who created it, how, and why. On the first, it asks whether it is "self-evident to your visitors who authored your content" and recommends that pages "carry a byline, where one might be expected", with bylines that "lead to further information about the author or authors involved." The same document is careful about what this is not. It says that "E-E-A-T itself isn't a specific ranking factor", and that rater data "is not used directly in our ranking algorithms."

The Search Quality Rater Guidelines are where the framework lives. Google added the extra E for experience in December 2022 and said then that, of the four elements, "trust is most important." The September 2025 edition added examples of how raters should assess AI Overviews and broadened the "Your Money or Your Life" category to cover civic information such as elections. None of this creates an author score. It describes what Google's systems are trying to reward: content from sources a reasonable person would trust on that topic.

Google is also explicit about the failure mode. Its guidance calls "fabricating creator profiles (such as by using AI-generated headshots, made-up names, or false credentials...)" a form of deception. A fake expert byline is worse than no byline.

Why a byline plausibly matters to AI engines anyway

There are two mechanisms worth taking seriously, even without a published effect size. First, the AI engines with web search do not rank the web from scratch. Google's AI Overviews are built on its own Search index and ranking systems, and studies of ChatGPT search find its citations track Bing's top results closely. Whatever quality signals those underlying systems use, including the signals Google says are designed to identify trustworthy content, are inherited by the answers built on them.

Second, a named author is an entity that can be resolved. A byline linked to an author page, which links to the person's professional profiles and is described in structured data, gives an engine a way to connect the page to a real person with a track record. An anonymous "editorial team" gives it nothing to connect. In regulated fields this goes beyond search: medical, legal, and financial content often needs a named, qualified reviewer for professional reasons, which our guide for medical practices covers in detail.

The author setup worth doing

None of this is expensive. For any site that publishes long-form content:

1. Put a real person's name at the top of every article, not "Admin" or "Our Team." If several people contributed, name each one, and name the reviewer separately for medical, legal, or financial topics.
2. Link the byline to an author page that explains who the person is, what they have done, and why they are qualified to write about the topic, with links to their professional profiles.
3. Mark up the article with Article or BlogPosting structured data and an author of type Person. Google's documentation asks for the name alone in author.name, a url or sameAs that identifies the person, and a separate author entry for each author.
4. Use the same name, photo, and job title everywhere the person appears: the site, LinkedIn, conference pages, and podcast show notes. Consistency is what lets an engine treat them as one entity.
5. Show a published date and a last-updated date, and only change the updated date when the content substantively changes.
6. Cite primary sources on the page. Authorship tells the reader who is making a claim; citations show them why to believe it.

What a useful author page contains

An author page is for the reader who wants to know whether to trust the writer, so write it for that reader. In practice that means a plain first sentence stating who the person is and what they do; the specific experience behind the topics they write about, such as years in practice, licenses, or the companies they have built; a short list of their published work on the site; links to their professional profiles and any third-party coverage; and a way to contact them or the editorial team. Mark the page up with a Person entity that uses the same name and job title as the bylines that link to it.

What not to buy

Several tools sell an "E-E-A-T score" for a page. Google has said E-E-A-T is not a specific ranking factor and that rater data is not used directly in ranking, so any single number that claims to measure it is the vendor's own model, not Google's. The same caution applies to any multiplier for how much more often authored pages are cited, including the one we used to publish, unless it comes with a method you can check. Read the glossary entries on E-E-A-T and author authority for the longer history.

How SEO for AI Agents measures this

We deliberately do not produce an E-E-A-T score or an authorship score, because we cannot tie either to a published method. What we report are the verifiable facts underneath. The JSON-LD validity check parses the structured data on your editorial pages and shows whether an Article block exists, whether it declares an author, and whether it is valid against the Schema.org vocabulary. The entity graph check shows whether your organization is anchored to the identity profiles knowledge graphs resolve against.

And for the question this article could not answer, whether authored pages actually get cited more, our source of citation check records the exact URLs each engine cites, alongside the verbatim answer, and the citation measurement reports how often you are cited with a confidence interval. Over time that is how we intend to replace a retracted multiplier with a real, reproducible one. Until then, add the byline because it is good practice, not because of a number.