In the answers generated by AI, "who said it" and "what was said" will be weighed together. Most Taiwanese B2B websites treat the author field as a plain text name. The AI engine only reads a few words, and there is no way to connect this content back to a real and verifiable expert. The role of person schema is to upgrade your founder or content author from a string of words into an entity with identity, expertise, and endorsement in the knowledge graph.
Why AI engines need to know who the author is
Google's E-E-A-T lists "experience" and "professionalism" as the core factors for evaluating content quality, but the machine does not automatically know who is behind an article. When ChatGPT, Perplexity, or Google AI Overviews decide which piece of content to cite, they favor pages with clear sources, identifiable authors, and alignment with external authoritative sources. Entity recognition (entity disambiguation) solves the problem of the same name: there are hundreds of people named "Chen Zhiming", and the engine needs enough signals to determine that Chen Zhiming in your article is the consultant who has written thirty articles in the SaaS field and works for a certain company. Person schema provides exactly this set of signals.
Minimum usable structure of Person schema
It is not necessary to fill all the fields at once. An author entity that can be correctly parsed by the engine usually contains the following attributes, written in JSON-LD on the author's personal page and article page:
- name: The full name of the author, which must be completely consistent between the entire site and the external platform. Even the spaces and English spelling must be consistent.
- @id: A stable set of URL-based identifiers (such as https://yoursite.com/author/name#person) that allow other schemas to refer back to the same entity.
- jobTitle and worksFor: job title and organization, worksFor points to the company's Organization entity.
- sameAs: Points to an external file array that can verify identity. This is the most critical column.
- knowsAbout: A list of topics that the author specializes in, corresponding to the content areas in which he has actually published.
- alumniOf, description, image: academic experience, introduction and headshot to enhance background and credibility.
The problem is that there is only one link in the environment, such as "author" if you don't bring this step, but the information can be very interesting.
sameAs: Connect the entity to the knowledge graph
sameAs is the column with the highest reporting rate in the entire set of structured data. It lists the file URLs of the same person on other authoritative platforms. The engine compares these links and can map the author on your website to the existing entity node in its knowledge base. In practice, priority is given to filling in these types of sources:
- LinkedIn profiles are the most important in B2B scenarios and are easiest to be trusted by engines.
- X/Twitter, Threads and other social accounts.
- Wikidata or Wikipedia entries, if any, are highly weighted.
- Founder page on Crunchbase, AngelList.
- Academic authors can add ORCID and Google Scholar.
- The company’s official website’s team page, speech event page, and media interviews and reports.
A reminder: sameAs emphasizes bidirectional consistency. If you claim that the author's LinkedIn is a certain URL in the schema, it is best that the LinkedIn profile can be linked back to your website or company, so that the signal is correct. Making a unilateral announcement but finding no such person on the other side will dilute the credibility.
Create an author's "physical homepage"
Each important author should have an exclusive personal page with a fixed URL and stable content, which serves as the official node (entity home) of this entity in your domain. This page carries the complete Person schema, real profile, links to masterpieces and external files. The author field of the article page then uses @id to point back to this page, forming a single source of truth. Without the physical homepage, author information will be scattered in the byline of each article, making it difficult for the engine to aggregate them into a stable entity.
Use @id to connect authors, articles and organizations into one picture
The power of structured data lies in the connections between nodes. The article uses the author of the Article schema to point to the author's @id; the author's worksFor points to the company's Organization @id; the company's Organization then uses employee or founder to point back to the author. In this way, what the engine reads is not three isolated pieces of information, but a small knowledge graph that corroborates each other. When it wants to judge "who is responsible for the content of this company and whether that person really understands this question," the answer can be seen in the picture.
When we take inventory of structured data for B2B customers, the most common vulnerability is not the lack of content, but the fact that the author of each article "cannot be found" - the engine reads the name, but cannot connect to any verifiable source. Supplementing Person schema and sameAs is usually the fastest step to see results.— Tenten GEO implementation experience
Landing Checklist
- Use the Google Complex Search Results Testing Tool and the Schema.org validator to verify that the Person markup is correct.
- The author's name, professional title, and company name are spelled exactly the same across the site, including Chinese and English, and spaces.
- sameAs Every link can be opened, and the opposite file can recognize your website or company.
- The topics listed in knowsAbout should be in line with the content areas that the author has actually published in, so don’t overdo it.
- The author of each article uses @id to point to the same author entity, instead of repeatedly posting a string.
- The author's entity homepage is indexable, not noindexed, and not blocked behind a login wall.
Start with an author: select the most representative founder or consultant in the company, build a physical homepage for him, fill in the Person schema and sameAs, and then point back the author fields of all related articles. After completing a complete example, other authors can copy it. If you want to know whether your content is "no author found" in the eyes of the AI engine, and what physical signals are missing, you can make an appointment for a 30-minute GEO diagnosis (/contact), and we will directly point out a few areas that can be fixed first.
