What is schema markup, in plain terms?
A web page is written for people. Schema markup adds a small, hidden layer written for machines. It says, for example: this page is about a Product, its name is Acme Payroll, it costs Rs 999 a month, and it's made by an Organization called Acme whose official profiles are these. The vocabulary comes from schema.org, which Google, Microsoft and others created together.
Does schema help with ChatGPT, Gemini and AI Overviews?
The official statements point in slightly different directions. At SMX Munich in 2025, Microsoft's Fabrice Canel said schema markup helps Microsoft's language models understand content, which matters for Bing and Copilot. Google's documentation on AI features takes a more careful line: you don't need to add special structured data to appear in AI Overviews or AI Mode, and structured data you do use should match what's visible on the page.
OpenAI, Anthropic and Perplexity haven't published guidance either way. What's clear is that their crawlers fetch your HTML, and JSON-LD sits in that HTML. A clean, accurate block of facts can't hurt their understanding, and it may help.
So the honest answer: schema is a supporting signal. It makes your facts easier to read and harder to misunderstand. It doesn't replace having those facts on the page, having crawlers allowed in, or being mentioned elsewhere.
Which schema types matter most?
| Organization | Home page | States your official name, logo, description and profiles (sameAs), so mentions elsewhere connect to one entity |
|---|---|---|
| Product or Service | Product and pricing pages | Labels what you sell, price and availability, which often come up in buyer questions |
| Article or BlogPosting | Blog posts and guides | Shows the author, publisher and dates, useful when assistants judge freshness and credibility |
| FAQPage | Pages with genuine Q&A | Marks question-and-answer pairs that match how people ask AI |
| BreadcrumbList | Every page | Explains where a page sits in your site, helping crawlers understand structure |
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Organization",
"name": "Acme Payroll",
"url": "https://acme.example",
"logo": "https://acme.example/logo.png",
"description": "Payroll and tax filing software for small businesses in India.",
"sameAs": [
"https://www.linkedin.com/company/acme-payroll",
"https://x.com/acmepayroll"
]
}
</script>Use the same description here that you use on your site, social profiles and directories.
Common schema mistakes that hurt more than help
- Marking up things that aren't on the page. FAQ schema for questions nobody can see, or prices that differ from the visible ones, is against Google's guidelines and confuses everyone.
- Stale values. An old price in your Product schema is a fact an assistant may repeat. Update schema when the page changes.
- Five different descriptions of your company. If your schema, About page and LinkedIn all describe you differently, you've made it harder, not easier, to understand who you are.
- Plugins that add everything. Some tools add dozens of types by default. More markup isn't better; accurate markup is.
- Never testing. Run pages through Google's Rich Results Test or the Schema Markup Validator after changes.
Where schema fits in your priorities
If you're short on time, do things in this order. Make sure AI crawlers can reach you (see robots.txt for AI crawlers). Make sure your key facts are written in plain text on the page. Add Organization and Product schema that match those facts. Then spend most of your effort on being mentioned in the places AI reads, which our guide on brand mentions vs backlinks covers.
Frequently asked questions
Does schema markup help ChatGPT understand my website?
OpenAI hasn't said either way. Its crawlers read your HTML, which includes JSON-LD, so accurate schema can make your facts clearer. Microsoft has said schema helps its own AI models.
Do I need schema to appear in Google AI Overviews?
No. Google says no special structured data is required to appear in AI Overviews or AI Mode, though it recommends that any structured data you use matches the visible content.
Which schema is most important for AI search?
Organization schema on your home page, with consistent sameAs links to your official profiles, is the best place to start. Product or Service schema on commercial pages comes next.
Is FAQ schema still worth adding?
Add it where you have genuine, visible questions and answers. Google now shows FAQ rich results for very few sites, but the clear question-and-answer format suits how people ask AI tools.