Back to Blog
  • Entity SEO
  • AI Search Visibility

Schema Markup for AI Search: What Google, Microsoft and ChatGPT Say

Hayden BondUpdated 9 min read
Schema Markup for AI Search: What Google, Microsoft and ChatGPT Say
Google says you don't need schema markup for AI search on Google, Microsoft says schema helps search engines and AI systems understand your content, and OpenAI doesn't say anything about it. So schema isn't the switch that gets you named in AI answers. I'd still add it, because it helps in regular search and it keeps your facts straight. Here's what each company says, and where I'd start on your site.

Not required

Google on structured data for its AI answers: "there's no special schema.org markup you need to add"

Source: Google Search Central, July 2026 ↗

What schema markup is, in plain words

When I was a kid, I spent a lot of time with my grandparents in a rural farming town. Every morning and evening, my grandfather sat at his desk and spoke with people around the world.
He was a HAM radio operator. With call signs and a UHF setup, he could reach Europe, Asia, or just the next county over. Long before the internet, he'd built his own network.
It only worked because the system was structured. Everyone had identifiers. Every exchange followed a protocol. If you didn't know the format, you couldn't make contact.
Schema markup is a format like that for your website. It's a small block of code that labels the facts on a page in a way software expects: this is a business, this is its name, this is its address, these are its hours. The labels come from Schema.org, a shared list of types that search engines read. Google recommends a format called JSON-LD, which sits in a script tag on the page. Your visitors never see it.
The difference from my grandfather's radio is that a search engine can still read your page without the code. Schema makes the facts easier to pick out. It doesn't decide whether anyone hears you.

Does Google need schema markup for its AI answers?

No. Google's guide to its AI features, updated July 10, 2026, says it plainly: "Structured data isn't required for generative AI search, and there's no special schema.org markup you need to add." Google's page on AI Overviews and AI Mode says specific optimization isn't required there either, and the usual SEO basics still apply.
In the same breath, Google says to keep using it, "as it helps with being eligible for rich results on Google Search." Rich results are the fancier listings in regular search, like star ratings, prices or recipe photos under a link.
Google has its own evidence for that part. In a 2018 case study, Google says Rakuten Recipe, which had used recipe structured data since 2012, expanded it and worked with Google on it in 2017. Traffic to its recipe pages from search engines went up 2.7 times.

2.7x

Rise in search engine traffic to Rakuten Recipe pages after it expanded its recipe structured data with Google. Regular search, 2018, not AI answers.

Source: Google Search Central case study, May 2018 ↗
That's a recipe site with a lot of pages, in regular search, eight years ago. I wouldn't expect those numbers for a local business. I share it because it's the proof Google itself points to.
Google also asks for one more thing that matters: your structured data has to match what's visible on the page. Schema that says you open at 7 while the page says 8 is worse than none, in my view.

What Microsoft says about schema and Bing's AI answers

Microsoft goes further than Google. In an October 2025 post on optimizing for AI search answers, Microsoft's Krishna Madhavan wrote that "Schema is a type of code that helps search engines and AI systems understand your content." Bing's index powers Copilot, so that's the closest any of these companies comes to saying schema helps its own AI understand your pages.
The same post has a line I'd tape to every small business owner's monitor: "Avoid putting key information only in images." Microsoft says AI systems can sometimes read text in images, but it "often reduces accuracy."
That's the first thing I'd look for on your site. A menu saved as a photo. Hours on a flyer. A license or a certification shown only as a badge. A person can read all of it. Software may not. Put those facts in plain text on the page, then add schema that says the same thing.

What OpenAI says about schema and ChatGPT

Nothing, as far as I can find. OpenAI's page about its crawlers covers which bots visit your site and how to let them in. It doesn't mention schema or structured data. The closest OpenAI comes is its shopping help page, which says ChatGPT considers "structured metadata from first-party and third-party providers (e.g., price, product description)" when it picks products to show. It doesn't say whether that means schema on your own pages.
The best outside evidence I know of is Kurt Fischman's February 2026 working paper. It looked at citations from ChatGPT and Gemini and found that pages that got cited used schema about as often as pages that didn't. Pages with Product or Review schema, with details like prices and ratings filled in, did better. I go through the numbers in how ChatGPT gets its information. It's one paper on older models, so I treat it as a lead.
My read: having schema isn't what gets you cited. Filled-in facts might help.

Can AI tools read schema that loads with JavaScript?

Some can't. In December 2024, Vercel looked at AI crawler traffic across its network and found that none of the major AI crawlers it tested ran JavaScript. That list included OpenAI's OAI-SearchBot, ChatGPT-User and GPTBot, Anthropic's ClaudeBot, and PerplexityBot. Gemini, which uses Google's crawler, did. That's Vercel's measurement of its own network, almost two years old, so things may have changed.
Google says it can read JSON-LD that's added to the page by JavaScript. So if your schema only shows up after a script runs, Google may see it while those other crawlers might not. Still, it's an easy fix. Ask whoever runs your site to put the schema in the page's HTML from the start.

Which local business schema to add first

If you have a storefront, an office or a service area, start with local business schema. Google's guide to it lists two required details, your business name and address, and recommends more, like hours, phone number, the web page for that location and a link to your menu if you serve food. It also says to use the most specific type you can, like Restaurant or HealthClub instead of plain LocalBusiness.
Here's a made-up example. Say you run a surf school. Your schema would give the school's name, street address, phone number and hours, and point to the page about lessons. The same facts would be in plain text on that page.
After that, I'd add:
Then check your work in Google's Rich Results Test. It only checks the rich result types Google supports, so it won't tell you how ChatGPT reads the page. It will catch broken code.

Say the same facts the same way everywhere

Schema on your site is one place your facts live. Your Google Business Profile, your directory listings and your social profiles are others. If your hours or service area say one thing in your schema and another on Yelp, you've told the internet two stories.
I think this matters more than any one block of code. My guess is that a model that's read the same name, address and description in many places has less to guess about. I covered the listings side in how to get your business found by AI.
For Hawaii businesses, I'd name your town and island in plain words, on the page and in the schema. I can't prove that helps AI answers. It costs you nothing, and it's what a visitor would type.
Plate Lunch Collective is an AI SEO agency that gets brands named when buyers ask, in search, on social, and in AI. Schema and consistent listings are part of our entity SEO work.

Where I'd start on your own site

You don't need all of this at once. Here's the order I'd do it in.
  • Find any key facts that only live in images and put them in text.
  • Add local business or Organization schema with your real name, address, phone and hours.
  • Make sure the schema matches what the page says.
  • Ask whoever runs your site whether the schema is in the HTML or added by a script.
  • Run the page through Google's Rich Results Test.
  • Check that your Google Business Profile and main listings say the same thing.
If you'd rather we take a look, the AI Search Game Plan shows where you stand in Google and in AI answers, who comes up instead, and what to do first.
My grandfather's radio only worked because he learned the protocols. Without call signs, he was shouting into static. Schema is one protocol worth learning for your website. It helps some tools read your facts clearly. Whether you get named, I think, still comes down to what's said about you, and whether it's said the same way everywhere.

Common questions

Is there special schema markup for AI search?

No. Google says there's no special schema.org markup to add for its AI features, and Microsoft points to the ordinary Schema.org types, like product, review and event. Schema markup for AI search is the same schema you'd add for regular search. I'd get it right, and I wouldn't count on it alone.

What is local business schema, and what goes in it?

Local business schema is code that tells search engines your business's name, address and other details in a standard format. Google requires the name and address, and recommends hours, phone number, the web page for that location and, for food businesses, a menu link. It also asks you to pick the most specific type that fits, like Restaurant instead of LocalBusiness.

Does schema markup help with Microsoft Copilot?

Microsoft says schema "helps search engines and AI systems understand your content," in its October 2025 guide to AI search answers, and Bing's index powers Copilot. It doesn't say schema gets a business named in Copilot's answers, or how much weight it carries. I'd treat it as one piece among several.

Is FAQ schema still worth adding?

Not for Google's search listings. Google stopped showing FAQ rich results on May 7, 2026, and removed the documentation in June. I'd still keep an FAQ on the page, written for people, because the questions and answers are useful in plain text.

Can I add schema without a developer?

Often, yes. Microsoft says schema is often added through your website platform, like WordPress, Shopify or Wix, or by a developer adding it to the page code. Check the result in Google's Rich Results Test either way.

How do I check that my schema works?

Paste your page's address into Google's Rich Results Test. It shows the schema it found and flags errors, but only for the rich result types Google supports. It can't tell you how ChatGPT or Copilot reads your page, so read the page yourself too, and make sure the facts in the schema match what's written there.

Sources

Share this article
Add as a preferred source on Google

Topics

Curious how your business looks on Google and in AI answers?

Plate Lunch Collective is an AI SEO agency that gets brands named when buyers ask, in search, on social, and in AI. If you'd like to talk it through, we're easy to reach.

Work with us

Hayden Bond

Hayden Bond, in SEO since 2004, founded Plate Lunch Collective, an AI SEO agency that gets brands named when buyers ask, in search, on social, and in AI.

Stay current on AI search

New research and analysis, direct to your inbox.

Related articles

Explore our Entity SEO services

Previous

Is SEO Dead? How to Tell If Your Site Broke or Search Changed

Next

Voice Search Optimization: What Changes When People Ask Out Loud