Author & Person Schema Depth: From Name String to Verified Expert

Author and Person schema depth evaluates how much verifiable detail your author markup provides to AI engines. The Tier 0 audit checks whether Person schema exists -a yes/no question. This intelligence-level evaluation goes far deeper: does the markup contain enough information for AI engines to independently verify the author's identity, credentials, and expertise?

AI engines use author information as a critical trust signal. Here's what Claude actually sees when it encounters an article about healthcare compliance with rich author markup: a Person with name, jobTitle "Healthcare Compliance Director," alumniOf "Johns Hopkins University," knowsAbout ["HIPAA", "Patient Rights", "Healthcare Law"], and sameAs links to LinkedIn and a professional association profile. It can cross-reference these claims. The author becomes a verified expert entity -not just a name string.

The evaluation assesses multiple dimensions. Identity verification -do sameAs links point to real, accessible profiles? Credential chains -do the stated jobTitle, alumniOf, and affiliation properties support expertise claims for the content topic? Body of work connectedness -is the author linked to other published content, enabling AI to assess their track record? Topic alignment -do the author's stated knowsAbout areas match the topics they write about on your site?

The scoring also accounts for consistency across pages. An author with different jobTitle values on different pages, or name variations preventing AI engines from connecting their articles, loses authority points. AI engines build entity models -they try to construct a unified picture of who this author is and what they know. Inconsistent markup fragments that model.

Checking whether Person schema exists is a binary test any automated tool can perform. Evaluating whether that schema actually enables AI trust verification requires the same kind of reasoning AI engines use. A Person schema with only a name property is technically valid but functionally useless -it gives the AI no ability to verify credentials or assess expertise.

AI-level testing matters because AI engines perform increasingly sophisticated author verification. When ChatGPT processes a medical article, it doesn't just check for an author name. It follows sameAs links to verify the person exists on LinkedIn, checks whether stated credentials match their profile, evaluates whether their other published work covers similar topics, and weights the article's credibility accordingly. A page claiming to be written by a "healthcare expert" with no verifiable identity gets treated as anonymous content.

We've found the citation rate difference between deep and shallow author schemas to be substantial in competitive niches. In the patient advocacy space, articles with full author schemas (name, title, credentials, profile links, knowsAbout) got cited 3-4 times more often than comparable articles with only author name attribution. The content quality was similar. But AI engines could verify one author and not the other.

This evaluation also catches a common problem: author schemas that are technically complete but factually unverifiable. If your Person schema includes a sameAs link to a LinkedIn profile that doesn't exist, or lists a jobTitle for a company with no web presence, the schema actively damages trust. AI engines that follow dead sameAs links learn to distrust the source. The Intelligence Report identifies these broken verification chains before they undermine your credibility.

The author schema depth analysis begins by extracting all Person schema instances from your site's pages. For each unique author entity, the system builds a profile from markup properties: name, jobTitle, worksFor, alumniOf, knowsAbout, sameAs, description, image, and any other Person properties.

First evaluation layer: schema completeness. A full Person schema for authorship should include at minimum -name, jobTitle, worksFor (Organization), at least one sameAs link, and relevant knowsAbout entries. The report scores completeness as a percentage and provides a checklist of missing properties.

Second layer -this is where AI-level testing comes in. The system follows sameAs links and verifies linked profiles exist and are accessible. It checks whether information on those profiles is consistent with schema claims. If your schema says "Director of Patient Advocacy" but their LinkedIn says "Marketing Coordinator," the inconsistency gets flagged. If the sameAs URL returns a 404, the broken link gets flagged as a trust-damaging signal.

Third layer: topic alignment. The system uses AI to analyze whether the author's stated knowsAbout areas, job title, and organization are relevant to the topics they write about on your site. An article about AI optimization credited to someone whose schema shows expertise in "culinary arts" creates a misalignment AI engines will detect.

Fourth layer: cross-page consistency. The system compares the author's Person schema across every page where they appear, flagging inconsistencies in name spelling, title, organization, or knowsAbout lists. It also checks whether the author has a dedicated author page consolidating their profile and linking to all their content -a pattern significantly strengthening the author entity model AI engines build.

The output includes a per-author report card: completeness score, verification status for each sameAs link, consistency assessment, topic alignment score, and specific recommendations for strengthening each author's schema.

Above 80: your author markup provides enough verifiable detail for AI engines to treat the author as a confirmed expert. This typically requires complete Person schema with at least two working sameAs links, accurate jobTitle and worksFor entries, relevant knowsAbout arrays, and consistency across all pages.

Between 50 and 80: partial author schemas providing some trust signals but leaving verification gaps. The usual culprits -missing sameAs links (the schema identifies the author but doesn't enable independent verification), absent knowsAbout properties (the AI can't confirm topic expertise), or inconsistent information across pages. These gaps are usually straightforward to fix. Add LinkedIn and Twitter URLs to sameAs. Populate knowsAbout with 5-10 relevant expertise areas. Standardize the schema across pages.

Below 50: author schemas either absent or too thin for meaningful trust signals. Many sites in this range use the author name in visible text but don't include Person schema at all, or include one with only the name property. From an AI trust perspective, these pages are effectively anonymous.

The verification status of sameAs links deserves special attention. A broken sameAs link is worse than no sameAs link at all. When an AI engine follows a sameAs URL and gets a 404, it learns your schema contains unverifiable claims. This doesn't just affect the author's credibility -it can impact trust signals for the entire site. The Intelligence Report lists every sameAs link and its verification status so you can fix or remove broken ones immediately.

For multi-author sites, compare author scores against each other and against the topics they write about. If your highest-credentialed author (schema depth 90) writes about general company news while your entry-level contributor (schema depth 30) writes your most important technical content -there's a strategic misalignment. Assign your strongest author entities to your most important content clusters.