Organization and Person Schema with sameAs: Build Your Entity

Organization and Person are schema.org types that tell machines who the company is and who the author is. The sameAs property links your site to profiles that identify the same entity, such as LinkedIn, YouTube or Wikidata, and @id gives each entity a unique identifier that you reuse throughout your JSON-LD. This helps disambiguate who you are. It does not guarantee a knowledge panel, a rich result or an AI citation, as the overview of SEO, AEO and GEO shows.

Quick summary

• An entity is something identifiable without ambiguity: a company, a person, a brand. Schema helps declare who is who.
• Google says there is no special schema for appearing in AI answers. The value lies in clarity and consistency, not in a shortcut.
• Use @id to create a single node for Organization, WebSite and Person, and reference those nodes in your articles instead of repeating data.
• sameAs should point only to pages that identify the same entity, such as official profiles or a Wikidata item.
• Wikidata and the knowledge panel depend on external criteria: you cannot create them by decree.
• Validating the JSON-LD and checking that it matches the page's visible content is part of the job.

How to read this: if you have never worked with schema, read "What is an entity?" and copy the example, swapping in your own data. If you already do technical SEO, go to the full @graph, the sameAs rules and the section on limits. This text complements the general guide to schema markup for AI Search.

What is an entity and why does it matter for AI?

An entity is something that exists distinctly and can be identified without confusion, such as "company X, with this website and this tax ID" or "author Maria Souza, SEO consultant." For search engines and assistants, names repeat and get mixed up. Declaring the entity with structured data reduces the chance of it being merged with another one of the same name.

Google itself describes this goal. According to the Organization structured data documentation, adding this data to the home page helps Google understand the organization's administrative details and tell it apart from others in results. Some properties influence visual elements, such as the logo shown. The documentation also says there are no required properties and that you should add the ones that apply.

In plain English: think of schema as the company's ID card on the internet. The name alone may be common, but name plus website, logo and official profiles make an identity that is hard to confuse. sameAs is like saying: "this same company also appears at these other addresses."

For the broader concept, see entities and semantic SEO.

What does Google say about schema in AI answers?

Google states that structured data is not required for generative AI search and that there is no special schema.org markup to add. It still recommends continuing to use structured data as part of your overall SEO strategy, because of eligibility for rich results. That is in the guide to optimizing for generative AI features.

As for other assistants (ChatGPT, Perplexity, Claude), I found no official documentation confirming that they read Organization or Person JSON-LD to decide citations. So the honest stance is: do it because it is good identity and SEO practice, not because anyone promised an effect on AI.

For SEO professionals: treat JSON-LD as a layer of disambiguation and consistency, not a ranking factor. The practical gain is usually indirect: less conflict between the brand name, the logo and the profiles that systems find, and organized data for whenever some consumer (Google or third parties) decides to use it.

Which Organization properties should you actually use?

Use few, and true ones: name, URL, logo, description and profiles. Google's documentation lists, among the Organization properties, name, alternateName, legalName, url, logo, sameAs, foundingDate, contactPoint and vatID. None is required.

Property	What it is for	Caution
name	Name of the organization	Use the same name and alternateName that appear on the site
legalName	Registered company name, if different from the name	Include it only if it really exists
url	The organization's website	Helps Google identify the organization uniquely
logo	Representative logo	Use a stable, crawlable image
sameAs	URL of another page with information about the organization, such as a social media profile or a review site	Can have several URLs, but only of real, official profiles
foundingDate	Founding date (ISO 8601)	Provide it only if it is public and correct
contactPoint	Contact channels	Use details that are already visible on the site

How do you structure everything with @id and @graph?

Give each entity a stable @id, reuse that @id wherever the entity is mentioned and group the nodes in a @graph. The @id works as the node's ID number inside your JSON-LD. A common convention is to use the page URL plus a fragment, such as https://www.exemplo.com.br/#organization. The address does not need to open a new page, it only needs to be unique and stay the same.

Hypothetical example: "Studio Exemplo" is a fictional SEO consultancy with one author, Maria Souza. The block below goes on the home page (or sitewide, via a template). The data is made up and written in Portuguese; swap in yours, in your own language.

{
  "@context": "https://schema.org",
  "@graph": [
    {
      "@type": "Organization",
      "@id": "https://www.exemplo.com.br/#organization",
      "name": "Studio Exemplo",
      "legalName": "Studio Exemplo Consultoria Ltda.",
      "alternateName": "Exemplo",
      "url": "https://www.exemplo.com.br/",
      "logo": "https://www.exemplo.com.br/img/logo.png",
      "description": "Consultoria de SEO para clínicas e escritórios de contabilidade.",
      "foundingDate": "2019-03-15",
      "sameAs": [
        "https://www.linkedin.com/company/studio-exemplo",
        "https://www.youtube.com/@studioexemplo"
      ]
    },
    {
      "@type": "WebSite",
      "@id": "https://www.exemplo.com.br/#website",
      "url": "https://www.exemplo.com.br/",
      "name": "Studio Exemplo",
      "inLanguage": "pt-BR",
      "publisher": { "@id": "https://www.exemplo.com.br/#organization" }
    },
    {
      "@type": "Person",
      "@id": "https://www.exemplo.com.br/autores/maria-souza#person",
      "name": "Maria Souza",
      "jobTitle": "Consultora de SEO",
      "url": "https://www.exemplo.com.br/autores/maria-souza",
      "worksFor": { "@id": "https://www.exemplo.com.br/#organization" },
      "sameAs": ["https://www.linkedin.com/in/maria-souza-exemplo"],
      "knowsAbout": ["SEO local", "Schema markup"]
    }
  ]
}

In articles, you do not repeat the company or author data. You just reference the nodes by @id:

{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "Como configurar o Perfil da Empresa no Google",
  "image": "https://www.exemplo.com.br/img/perfil-empresa.webp",
  "datePublished": "2026-10-01T09:00:00-03:00",
  "dateModified": "2026-10-04T09:00:00-03:00",
  "author": { "@id": "https://www.exemplo.com.br/autores/maria-souza#person" },
  "publisher": { "@id": "https://www.exemplo.com.br/#organization" }
}

For SEO professionals: Google strongly recommends author in Article with type and url (or sameAs). As a precaution, an alternative to referencing by @id alone is to declare author directly in the article (@type, name and url) and leave the full node on the author's page. That is a safety choice of mine, not a documented requirement. Always test on the rendered HTML, because WordPress themes and plugins can generate several conflicting Organization nodes.

How do you mark up the Person (the author) without overdoing it?

Mark up the author with @type: Person, a real name, a URL that identifies them and, if they exist, profiles in sameAs. According to Google's Article documentation, in author.url you can use a page that uniquely identifies the author, such as a social media profile, an "about" page or a bio page. Google also accepts sameAs as an alternative and says it understands both when disambiguating authors.

The same documentation includes a rule that prevents common mistakes: author.name takes only the person's name, with no job title, no company name and no phrases like "by." The title goes in jobTitle. In practice, point the author URL to a real page with a bio, credentials and work. The markup only makes sense if the page exists and shows who the person is, which depends on authorship and trust (E-E-A-T).

How do you use sameAs correctly?

Use sameAs only for pages that identify the same entity without ambiguity. Schema.org defines the property as the URL of a reference page that unambiguously indicates the item's identity, for example the Wikipedia page, the Wikidata entry or the official website. Google documents it, for Organization, as the URL of a page on another site with more information, such as a social media profile or review site, and accepts several.

• In: active official profiles (LinkedIn, YouTube, Instagram, company page), the Wikidata item, a verified profile on review platforms that belongs to the company itself.
• Out: profiles of another person or company, generic directories, pages that mention you in passing, abandoned or duplicate profiles.
• Consistency: the name, logo and description on these profiles should match the site. Divergence gets in the way of identification.

What about Wikidata and the Knowledge Graph?

A Wikidata item is a public, structured record of an entity, and sameAs can point to it. But it is not a free-for-all registry. Wikidata's notability rules accept an item if it meets at least one of three criteria: it links to a page on a Wikimedia project, it refers to a clearly identifiable entity that can be described using public, serious references, or it fulfills a structural need.

For the knowledge panel, Google's help page explains that the information is generated automatically from public information on the web, that you can verify your identity and suggest edits, and that the current policy does not manually create or delete panels. In short: markup helps make the information readable, but it does not oblige Google to show a panel.

In plain English: Wikidata is a public record kept by the community. You only get in if public, serious sources talk about you. Registering yourself with no references is pointless, and an item created just to "gain authority" can be removed for not meeting the criteria. The path is to be genuinely mentioned, as we discuss in brand mentions without links.

Step by step to implement

Follow the order below and validate at each stage.

1. List the public facts: name, alternate name, registered company name (if any), URL, logo, description and official profiles.
2. Define the @ids for Organization, WebSite and each author, and keep them forever.
3. Publish the Organization and WebSite on the home page (or in a global template), preferably in a single @graph.
4. Create one page per author, with a real bio, and mark up the Person on it.
5. Reference the nodes in articles by @id or with the minimum author fields.
6. Validate in Google's Rich Results Test and at validator.schema.org, and check the rendered HTML with Search Console's URL inspection.
7. Compare with the visible content: the name, logo, contact details and profiles in the JSON-LD should appear on the site or be verifiable there.
8. Monitor whether the names appear correctly. Contentor's AI Visibility tracks, on the questions you choose, whether ChatGPT cites the brand (ChatGPT only), which helps you see whether the identity is being recognized. For the full diagnosis, use the step-by-step GEO audit.

Which mistakes and limits should you avoid?

The most common mistakes are treating schema as a shortcut and marking up things that do not exist. The table summarizes them.

Mistake	Why it is a problem
Two or three different Organization nodes on the same page	Conflicting information; check plugins and theme
sameAs pointing to profiles that are not yours	Mixes identities
A fictional author or an empty author page	Does not support trust and contradicts what the markup promises
Expecting a knowledge panel or citation just for having schema	Google does not guarantee a panel and does not require schema for AI
Outdated data (logo, name, contacts)	Creates divergence between site, profiles and JSON-LD
Copying an example without swapping the data	Publishes false information about your company

Frequently asked questions

What is sameAs in schema?

It is the property that gives the URL of another page that identifies the same entity, such as an official LinkedIn profile or a Wikidata item. It is used for disambiguation. The schema.org definition speaks of a reference page that unambiguously indicates the item's identity. There can be several URLs, as long as they truly belong to the same entity.

Does Organization schema help you appear in ChatGPT?

There is no official documentation confirming it. Google says there is no special schema for getting into AI answers. What the markup does is state clearly who you are, which is good practice. A citation depends on useful content, crawler access and external mentions, and is never guaranteed.

Do I need Wikidata to have Organization schema?

No. Wikidata is optional and only makes sense if your company meets the notability criteria, for example by having public, serious references. Many companies have a valid Organization with just a name, URL, logo and official profiles. Never create an item just to inflate authority.

Can I use the same entity across several sites in a group?

You can, as long as each site declares correctly who is who and uses consistent identifiers. In general, the parent organization and each brand can be different nodes, linked by parentOrganization (the schema.org property for the larger organization that one belongs to). Avoid copying the same Organization onto sites that represent distinct businesses, so you do not create conflicting information.

How do I know whether the schema was read correctly?

Validate in the Rich Results Test and at validator.schema.org and use Search Console's URL inspection to see how Google views the page. Remember that being valid does not mean being used: Google decides when and how to display any information.

Next step

If your site does not yet have a decent author page, start there: without a real person behind it, the markup supports nothing. Then implement the @graph above and review the rest in schema markup for AI Search. To connect this work to reputation off the site, see brand mentions without links and the step-by-step GEO audit.

Read also:

• Entities and semantic SEO
• Authorship and trust: E-E-A-T for AI
• SEO, AEO and GEO glossary