Structured data, entities and the Knowledge Graph for Gemini

#01What does Google actually say?

#02Which markup is worth having?

An Organization example:

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "Ledgerline",
  "url": "https://www.ledgerline.example",
  "logo": "https://www.ledgerline.example/logo.png",
  "description": "Accounts payable automation software for mid-sized finance teams.",
  "sameAs": [
    "https://www.linkedin.com/company/ledgerline-example",
    "https://www.g2.com/products/ledgerline-example",
    "https://github.com/ledgerline-example"
  ],
  "contactPoint": {
    "@type": "ContactPoint",
    "contactType": "sales",
    "email": "sales@ledgerline.example"
  }
}

Deeper guides: Organization schema for brand entities, SoftwareApplication schema for SaaS and Entity SEO: making your brand unambiguous.

#03What else makes your entity clear?

Markup is only one signal. Google's systems also read what the web says about you.

#04How do you check what Gemini believes about you?

Run these prompts in the Gemini app (logged out of connected apps, or in a clean account) and in AI Mode:

What is [Brand]? What category is it in, who is it for, and what does it cost?
Who are [Brand]'s main competitors?
Is [Brand] a good fit for [segment]? What are its limitations?

Log three things: wrong facts, missing facts, and the sources cited. Then:

1. Fix your own pages first (About, pricing, product).
2. Update third-party profiles that are cited with old information.
3. Check that the structured data matches the corrected text.
4. Re-run monthly (lesson 20).

#05How do you test markup?

• Rich Results Test for each template (homepage, product, article).
• Schema Markup Validator for general schema.org validity.
• Search Console enhancements reports for errors at scale.
• URL Inspection: check that the rendered HTML includes the JSON-LD if it's injected with JavaScript.

Our view: the most common problem we see isn't missing markup. It's contradictory markup, such as an old price in an Offer, an aggregateRating with no visible reviews, or three different company descriptions across templates. Fix contradictions before you add anything new.

#06Worked example: Ledgerline

Ledgerline is a fictional accounts-payable software company. Details are illustrative.

Gemini described Ledgerline as "an expense management tool", which is the wrong category.

Walkthrough

1. The SEO lead ran the three entity prompts in Gemini and AI Mode and logged the cited sources.
2. Crunchbase and the homepage slogan were the likely causes of the wrong category.
3. She wrote the entity statement and applied it across the site, Crunchbase, LinkedIn and G2.
4. The developer fixed the SoftwareApplication markup and validated it in the Rich Results Test.
5. Two months later both surfaces called Ledgerline "accounts payable automation software". The team noted the timing, without claiming the markup alone caused it.

#07Common mistakes

• Buying "AI schema" that Google says it doesn't need.
• Markup that doesn't match the page.
• Fake or unsupported aggregateRating.
• Slogans instead of a category in descriptions.
• Ignoring third-party profiles that AI cites.

#08Try it yourself

Ask Gemini and AI Mode "What is [your brand]?" Note the category and the sources. Then check your homepage Organization markup and About page. Do they state the same category in plain words? Write your 40-word entity statement and list every profile where it should appear.

#09Quick quiz

1. Is special schema needed for AI Overviews?

Answer: No. Google says no special schema.org markup is needed.

2. What must structured data match?

Answer: The visible text on the page.

3. Which property links your official profiles?

Answer: sameAs in Organization markup.

#10What should you check before the next lesson?

• Organization, product and Article markup validates and matches the pages.
• Your entity statement is consistent across your site and profiles.
• You've logged what Gemini and AI Mode say about you.

Previous: Google-Extended, Googlebot and the Search generative AI control. Next: Content formats that get cited in Gemini and AI Mode. Or return to the Gemini for Marketers track. Related: Entity SEO for AI, Schema markup for AI search, Knowledge graph.

#11Common questions

Is there special schema for AI Overviews or AI Mode? No. Google says there is no special schema.org markup needed. Structured data still helps rich result eligibility and must match the visible text on the page.

Which structured data types matter most for a brand? Organization (with logo, url and sameAs) on the homepage, Product or SoftwareApplication on product pages, and Article with author details on content.

Do llms.txt files help in Google AI features? No. Google Search says it ignores llms.txt and similar AI text files. They neither help nor harm Google visibility, though other services may use them.