Does an Author Name and a Credential Help AI Trust Your Site? What Google Says, What the Studies Don't Measure, and Where Our Audit Guesses
Two sites, two mistakes
Picture two small business sites.
The first belongs to a dental practice. Somebody read that credentials build trust, so the blog posts now carry a byline: "Dr. Alan Whitfield, DDS." There is no Dr. Whitfield. The name came out of a content tool, the headshot came out of an image generator, and the practice owner signed off on it because it looked professional.
The second belongs to a plumber. Twenty years in business, fully licensed, every customer on the route knows his name. And his name appears nowhere on his website. Not on the homepage, not on the About page, not in a footer. The site could belong to anybody.
One of these is a much bigger problem than the other, and it is not the one most people guess. Google's own guidance on helpful content warns against "Fabricating creator profiles (such as by using AI-generated headshots, made-up names, or false credentials." That is the dentist. The plumber has a missing feature. The dentist has a liability.
This post is about the space between those two. What should a small business actually say about who is behind the site, what does Google really document about it, and what does anybody actually know about whether AI answer engines care? The answer to that last one is shorter than the internet would like you to believe.
What Google actually documents
Start with the primary source, because a lot of advice about this topic is a game of telephone.
Google's guidance on creating helpful content says of the E-E-A-T family (experience, expertise, authoritativeness, trust): "Of these aspects, trust is most important." For topics where bad information could hurt someone's health, finances, or safety, which Google calls YMYL ("your money or your life"), the guidance says the content "must be highly accurate and consistent with established expert consensus." Google's systems "give even more weight to content that aligns with strong E-E-A-T" on those topics.
On authorship specifically: "We strongly encourage adding accurate authorship information, such as bylines to content where readers might expect it." Google also asks whether a byline leads "to further information about the author or authors involved." So the ask is a name, plus somewhere a curious reader can go to learn who that person is.
Now the sentence people skip. Google states that "E-E-A-T itself isn't a specific ranking factor." It is a way of describing what Google's systems try to reward, not a dial inside the algorithm.
None of that is a claim about AI answer engines. It is guidance about Google Search, written for site owners, and the rest of it is just as plain. Google's page on AI features says "There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary." It adds that "There's also no special schema.org structured data that you need to add," and that structured data should match the visible text on the page. It also says "You don't need to create new machine readable files, AI text files, or markup to appear in these features."
That last one covers Google's own AI features. It says nothing about ChatGPT, Claude, or Perplexity. What it does tell you is that even the company that wrote the E-E-A-T concept does not claim a byline unlocks a special AI door.
Raters are not the algorithm, but they show what "trust" means
Google also publishes a long document called the Search Quality Evaluator Guidelines. This is the instruction manual for the human raters who judge search results. I am quoting the September 11, 2025 edition, since the document gets revised.
These are instructions for people. They are not ranking signals. The ratings are not fed directly into the algorithm and the raters have no control over where anything ranks. What the document is useful for is seeing what Google means when it says "trust," because it spells it out in plain language.
The raters are told to work out who is responsible for a site and for its content. And here is the part that matters to a plumber: for a business homepage, the guidelines say "the website itself can be viewed as the primary content creator." A business site does not need a byline on its homepage to satisfy this. The business is the author. What raters want is for it to be clear who the business is and how to reach it. The guidelines also acknowledge that anonymity is legitimate on some kinds of sites, personal pages and user-generated content among them, and say that contact information matters most on sites that handle money or sensitive data.
On the trust question itself, the guidelines put it this way: "Trust is the most important member of the E-E-A-T family because untrustworthy pages have low E-E-A-T no matter how Experienced, Expert, or Authoritative they may seem." You can have a wall of credentials and still be rated low if the page is not trustworthy. Credentials are not a shortcut around that.
Expertise gets a concrete example in the same section: advice about electrical rewiring from a skilled electrician versus from an enthusiast of antique homes. Raters weigh what the site says about itself, what others say about it, and what is on the page.
Then the stakes change when the topic does. For clear YMYL advice pages, the guidelines say "a high level of expertise may be required for the page to be trustworthy." First-hand experience still counts, but some information "must come from experts." And for pages where trust matters most, "information about who created the content and who is responsible for the content is critical." The harshest line about missing information applies to a narrow case: YMYL pages or sites that handle sensitive data with absolutely no information about the website or content creator should be rated Lowest.
It does not say that a missing byline sinks a site. It says that a high-stakes page with no information at all about who is behind it is a problem.
Put the two businesses back next to each other. The plumber's homepage is not YMYL advice. The business is the content creator, and a name, a license number, and a phone number on the page would be easy to add and cost nothing. The dentist's pages about whether a symptom needs treatment are exactly the pages where "who wrote this and are they qualified" is a fair question. That is why one fake credential on a dental site is so much worse than a missing name on a plumbing site.
What the studies do and don't show
Here is where the confident advice usually shows up. Somewhere online you have probably seen a number attached to credentialed bylines, a percentage lift in AI citations or visibility, sometimes a correlation coefficient with decimals that make it look rigorous.
I went looking for the source of those numbers. What I found is a genre: vendor blogs, commercially motivated, unverified, and not in agreement with one another. I am not going to repeat the numbers, because repeating them would give them a credibility they have not earned.
The more careful material says something quieter. One roundup covering six studies on how AI engines cite the web describes none of them as measuring bylines, credentials, E-E-A-T, schema, or domain authority as citation factors. A Toronto paper (Chen, Wang, Chen, and Koudas, revised in May 2026) compares the kinds of source domains AI engines draw on, which is interesting, but it does not tell you that a name and a credential on your page changes your odds. I did not verify its figures.
Google documents why authorship and trust matter for Google. Nobody has shown, in anything I could find, that AI answer engines pick sources because of bylines or credentials. Absence of evidence is not evidence of absence, and it is entirely possible engines lean on these signals and nobody has measured it. But a claim with no measurement behind it is a guess, and a guess with a percentage stapled to it is still a guess.
If you add an author, here are the mechanics
Say you decide to name the person behind the content anyway. Good reasons exist that have nothing to do with AI. Customers want to know who they are hiring, and Google asks for it. Here is how to do it so the machine-readable version says the same thing as the visible one.
Google's Article structured data documentation is specific. The author.name field holds the name and only the name. A job title goes in jobTitle. Honorifics go in honorificPrefix and honorificSuffix. Use the Person type for a person and Organization for an organization. If there is more than one author, list each one separately.
The documentation also wants the author to be identifiable. author.url should link to a page that uniquely identifies the author, such as a bio page on your own site. sameAs works as an alternative, pointing to a page about that person elsewhere on the web. The Organization documentation uses sameAs the same way for a business: a URL on another website with more information about it. The purpose is disambiguation, so that "Whitfield Dental" and some other business with the same name do not get blurred together. It is a way of saying which one you are, not a way of ranking higher.
The most common mistake here is writing the author as a bare string, just a name in quotation marks, rather than a Person object with its own properties. The string works, in the sense that nothing breaks. It also tells a machine nothing beyond a name. If the mechanics of getting a block onto your page are what stopped you before, the platform-by-platform guide covers where it goes on WordPress, Wix, Squarespace, and GoDaddy.
One more rule of thumb, and it is Google's, not mine: structured data should match what a visitor can see. If the markup names an author the page never mentions, you have built the same problem as the fake dentist, just in a place fewer people look.
Where it stops being about rankings
There is a reason to be careful with health content that has nothing to do with search positions, and it comes from a study published in Nature Communications in April 2025. Wu and colleagues built a framework called SourceCheckup and tested seven large language models on 800 medical questions, which produced roughly 58,000 pairs of a statement and the source it cited. Their finding: "Between 50% and 90% of LLM responses are not fully supported, and sometimes contradicted, by the sources they cite." For GPT-4o with web search, "approximately 30% of individual statements are unsupported."
That study is not about bylines. It does not test whether credentialed authors get cited more. What it does say is that an AI answer about a health topic can misstate the page it points to. That has a practical consequence for a clinic. The page should be clear, specific, and easy to quote correctly, because the version of it that reaches a patient may have passed through a system that is not careful. A named, qualified clinician on a page about symptoms is still a good idea for the patient who lands there. That is a reason about patients, and it holds whether or not any engine rewards it.
What our audit does, and where it is guessing
Entity & Trust is worth 20 of the 100 points in an AEO Checker report. It runs five checks: business identity (do the name, address, and phone show up both in your schema and on the page), sameAs links, a byline, author schema, and an About page. The point split depends on which profile your site gets.
| Profile | Identity | sameAs | Byline | Author schema | About |
|---|---|---|---|---|---|
| Standard | 5 | 5 | 4 | 3 | 3 |
| Medical, legal, financial | 4 | 4 | 4 | 5 | 3 |
| E-commerce, real-estate listing, SaaS | 6 | 6 | 2 | 2 | 4 |
Standard covers every other kind of business, plus personal sites, organizations, and podcasts. For SaaS, the 6-point identity row is replaced by a check for pricing, legal pages, team information, and review profiles. Half credit rounds down, so a Standard-profile half on author schema is 1 point, not 1.5.
How each check works is mechanical, which is both its strength and its limit.
The byline check passes on a structural signal: a rel="author" link, or an element whose class contains words like author, byline, post-author, or entry-author. A plain-text lowercase "by" followed by a two- or three-word capitalized name, in the opening stretch of the page's paragraph text, earns half credit. The author schema check reads only the author of Article, BlogPosting, and NewsArticle markup. A standalone Person block elsewhere on the page does not count toward it. A Person-object author with a credential, or a detected byline with a credential nearby, earns full credit. A bare-string author, an Organization-typed author, or a byline without a credential earns half.
The "credential" part is a keyword match. The checker looks for words like MD, DDS, DVM, JD, CPA, PE, RN, PhD, "licensed," "certified," "board-certified," "years of experience," and phrases like "established 1998" or "since 2005," either in the schema author's hasCredential or jobTitle, or within about 200 characters of the byline name. That means it does not verify anything. It cannot tell a real DDS from an invented one, which makes the fake Dr. Whitfield from the top of this post a problem the audit would not catch. It also means a plumber's "licensed and insured, since 2005" counts.
The About check depends on the scan. The free Grade Check looks at one page, and on that run it gives a fixed 2 points as a warning if your homepage links to an About page. On the paid tiers, which crawl up to 20, 50, or 100 pages, it passes only if a crawled page has "about" in its URL. Otherwise it scores zero.
Now the guessing. The weighting for medical, legal, and financial sites, with author schema worth 5 points instead of 3, is a judgment call. It rests on Google's YMYL guidance and on the fact that adding an honest author and credential is cheap. It does not rest on a study showing AI engines prefer credentialed authors for medical content.
So what does the score tell you? That the signals Google describes are present in a form a machine can read. It does not tell you whether an AI engine will cite you for having them.
A ten-minute check
Open your About page and one service page, then view the source on each. Is there a named owner or author anywhere a visitor can see, for a business where people hire a person? If your site publishes articles, check that the author in the article markup is a Person object and not just a string. That applies to posts and articles. A plumber's homepage does not need Article schema, and I would not go adding it to satisfy a checklist.
If a credential appears anywhere, confirm it is real and that someone could verify it. License number, state board, whatever applies to your trade. If you cannot back it up, take it off the page.
Last, compare your About and contact details against your homepage. Same name, same address, same phone. A mismatch there is easy to miss and worth checking before you worry about a byline, and it is a thirty-second fix.
If you want a baseline
The free Grade Check scores your homepage from 0 to 100 and gives you a letter grade in about a minute, no card. It is a single overall number from one page, so it will not break out the author or About checks for you. The ten-minute check above is the part it cannot do for you.