Schema markup for AI search: what helps and what does not
Published 8 min read
Google statements checked on 11.11.2026. nenn.io markup read from the live pages and from src/lib/seo/schema.ts on 11.11.2026.
Does schema markup help with ChatGPT and Google AI? Google says its AI features need no special markup. Schema helps machines read facts you already show, like your name, price and FAQ.
Card needed. Cancel before day 7 and nothing is charged.
On the page: Fernhook Books
In JSON-LD: "name": "Fernhook Books"
On the page: Basic plan: 29 per month
In JSON-LD: "price": 29
On the page: Find us on LinkedIn
In JSON-LD: "sameAs": ["https://www.linkedin.com/company/fernhook-example"]The short answer
Schema markup is not a switch for AI citations in ChatGPT or Google AI. Google says AI Overviews need no special schema. Markup helps machines read facts you already show, so keep it equal to the visible text.
Google statements checked on 11.11.2026. nenn.io markup read from the live pages and from src/lib/seo/schema.ts on 11.11.2026.
Key points
Facts, not tricks. Markup repeats what the page says. It cannot add what the page lacks.
Must match the page. Google says not to mark up content readers cannot see.
No promise of citations. We found no public proof that schema alone makes AI name you.
What does Google say about schema for AI Overviews?
Google says there is no special schema.org markup to add for AI Overviews or AI Mode. A page must be indexed and able to show a snippet in Search.
- Google also says you need no new machine-readable files or AI text files to appear there.
- It lists one schema habit as good practice: structured data should match the visible text on the page.
- Other AI engines read your HTML text, so clear text on the page comes first.
- Google's full list for AI Overviews is in how to show up in Google AI Overviews.
Which 4 schema types matter for a company?
Four types cover most company sites. Each one states facts your pages already show.
| Type | What it says | Where it goes | Visible fact it must match |
|---|---|---|---|
| Organization (with sameAs) | Who you are, your logo and your profiles on other sites | Home page or about page | Your company name and the profile links on your site |
| Product or SoftwareApplication (with Offer) | What you sell and at what price | Product or pricing page | The price and plan names on that page |
| FAQPage | Questions and answers on the page | Pages with a visible FAQ | The same questions and answers, word for word |
| BreadcrumbList | Where the page sits on your site | Every page below the home page | The breadcrumb trail people see |
- Google says Organization markup helps it tell your company apart in search results.
- sameAs lists pages on other websites about your company, such as LinkedIn. That helps when another company has a similar name.
If AI mixes you up with another company, see when AI confuses your company with another.
What does Organization schema with sameAs look like?
Add one Organization block to your home page, or once for the whole site. Use the exact name, address and links your site shows.
{
"@context": "https://schema.org",
"@type": "Organization",
"@id": "https://fernhook.example/#organization",
"name": "Fernhook Books",
"url": "https://fernhook.example/",
"logo": "https://fernhook.example/logo-512.png",
"sameAs": [
"https://www.linkedin.com/company/fernhook-example",
"https://www.youtube.com/@fernhook-example"
]
}- List only profiles that exist and that you run. A dead link in sameAs claims a profile you do not have.
- Use a fixed @id. Other blocks can then point to the same company.
How should price and FAQ markup match your page?
Copy the numbers and words from the page into the markup, never the other way round. If the page says 29 per month, the markup says 29.
{
"@context": "https://schema.org",
"@type": "SoftwareApplication",
"name": "Fernhook Books",
"applicationCategory": "BusinessApplication",
"operatingSystem": "Web",
"offers": {
"@type": "Offer",
"name": "Basic, monthly",
"price": 29,
"priceCurrency": "EUR",
"url": "https://fernhook.example/pricing"
}
}- Best: build the markup from the same data as the visible page. Then a price change updates both.
- Mark up only plans you sell today. An old yearly plan in the markup is a claim about an offer that does not exist.
What should you not do with schema for AI?
Do not use markup to say things the page does not say. Google's rules are clear on this.
Do
Do: Mark up facts that readers can see on that page.
Do: Keep name, address and prices the same in markup, page and profiles.
Do: Test with Google's Rich Results Test and the Schema Markup Validator.
Do not
Avoid: Mark up text that is not on the page. Google says not to.
Avoid: Add ratings or reviews you cannot show on the page.
Avoid: Expect schema alone to make AI cite or name you.
- Google says structured data must be a true picture of the page content.
What schema does nenn.io use, and why?
We read our live pages and our markup code on 11 November 2026. Each block repeats facts the page shows, and each type comes from one shared code file.
| Page | Types | Why |
|---|---|---|
| Every page | Organization, WebSite | One identity with 5 profile links in sameAs. No site search box, so no SearchAction. |
| Home page | SoftwareApplication, FAQPage | The product without prices. The FAQ is built from the same strings people read. |
| /pricing | SoftwareApplication with 6 Offers, one per plan, FAQPage | Prices only where prices are visible, and only the monthly plans we sell. |
| Blog posts (from the code, not live yet) | BlogPosting, BreadcrumbList, FAQPage | No author markup until the author box is signed. Breadcrumbs come from the visible trail. |
How does Nenn check your schema and facts?
The Website check in the trial reads your titles, schema, sitemap and llms.txt. It also checks that your name, address and prices match across your pages.
Structure checks
Titles, schema, sitemap, hreflang and llms.txt, failed checks first.
Same facts everywhere
Name, address and prices compared across your pages.
Fix drafts
Nenn drafts the fix. You approve it before anything changes.
Questions about schema and AI search
Does schema markup help with ChatGPT?
Not as a switch. ChatGPT search reads the pages it finds, and clear text on the page matters most. Markup can repeat facts in a form machines read easily. We found no public proof that it makes ChatGPT name you.
Is there special schema for Google AI Overviews?
No. Google says there is no special schema.org markup to add for AI Overviews or AI Mode. The page must be indexed and able to show a snippet.
Should I still add FAQ schema?
You can, if the FAQ is visible on the page. Google no longer shows FAQ rich results in Search since May 2026. The markup still describes your questions and answers in a standard way.
What is sameAs in Organization schema?
A list of links to pages about your company on other websites, such as LinkedIn or YouTube. It helps search engines tell your company apart from others with a similar name.
Is JSON-LD better than llms.txt for AI?
They do different jobs. JSON-LD sits in each page and repeats its facts. llms.txt is one file that points AI tools to your key pages. Google says neither is needed for its AI features.
See if your facts are the same everywhere
The Website check in the trial reads your titles, schema, sitemap and llms.txt. Nenn drafts the fix for you to approve.
Card needed. Cancel before day 7 and nothing is charged.
What does Fernhook Books cost per month?
- 1Your price?
Keep reading
From correct markup to correct answers about your company.
When AI says something wrong about you
Find the source of a wrong claim and fix it.
Blog post
When AI confuses your company with another
Keep your name apart from similar ones.
Blog post
What is llms.txt?
The file, what it does and what research found.
Glossary
AI content writer
Nenn drafts the page or fix. You approve it.
Feature