TL;DR
- Schema markup doesn’t directly earn AI citations, based on the controlled tests published so far. In an Ahrefs study of 1,885 pages that added schema (May 2026), citations changed by -4.6% in AI Overviews and by amounts “statistically indistinguishable from zero” in AI Mode and ChatGPT.
- Google says you don’t need it for AI features. Its AI features documentation states there’s “no special schema.org structured data that you need to add.”
- Microsoft is the exception. Bing’s Fabrice Canel confirmed in March 2025 that schema markup helps Microsoft’s LLMs understand content.
- Google removed FAQ rich results on May 7, 2026, according to its documentation updates log, so FAQ schema no longer changes how your listing looks in Google.
- Schema still earns its place for entity clarity (Organization, LocalBusiness, Person), product data, and the rich results Google still shows. Rankings and content matter far more for AI citations.
What Is Schema Markup, and How Is It Different From Structured Data?
Schema markup is code that labels the facts on a page, such as your business name, address, author, or product price, using the shared schema.org vocabulary so machines can read them without guessing. Structured data is the general practice of labeling content this way, and JSON-LD is the format Google recommends for it.
In practice, the three terms describe one thing from different angles:
| Term | What it means |
|---|---|
| Structured data | Any standardized way of labeling page content for machines |
| Schema.org | The shared vocabulary of types (LocalBusiness, Person, Product) and properties (address, sameAs, price) |
| JSON-LD | The code format, usually a script block in the page, that Google recommends |
Google says it uses structured data “to understand the content of the page, as well as to gather information about the web and the world in general.” That second purpose, understanding entities, is where schema still matters most in an AI-heavy search world.
Does Schema Markup Help You Get Cited by ChatGPT, Perplexity, or Google AI?
No, not directly. Every controlled test published between late 2025 and mid-2026 found that adding schema did not measurably increase citations in AI answers. The studies that show schema “working” compare pages that already had schema with pages that didn’t, which mixes schema up with site quality.
| Study | What they did | Result |
|---|---|---|
| Ahrefs (May 2026) | 1,885 pages added JSON-LD, compared with 4,000 matched control pages | AI Overviews -4.6%; AI Mode +2.4% and ChatGPT +2.2%, both “statistically indistinguishable from zero” |
| OtterlyAI (March 2026) | Added five schema types to its own site, tracked 319 prompts on 7 AI platforms for three months | Schema “did not appear to directly influence AI Search citation behavior in a measurable way” |
| Search Atlas (December 2025) | Compared LLM visibility across schema categories | “No measurable effect on LLM visibility across OpenAI, Gemini, or Perplexity” |
| AirOps (April 2026) | Correlation across 353,799 pages | Pages with JSON-LD were cited 38.5% of the time vs 32.0% without, with no control for rankings |
Ahrefs explains that correlation gap itself. It found 53% of AI-cited pages have schema, but notes that schema “tends to live on better-maintained, more technically sophisticated sites.” The well-run site gets cited, and it also happens to have schema.
The Ahrefs study has limits worth knowing. A methodology critique by Misha Manko points out that it only covered pages already earning heavy AI Overview citations, pooled all schema types together, and measured a 30-day window. So the fair summary is “no proven lift,” not “proven useless.”
What does move citations? Search visibility. Ahrefs found 76% of AI Overview citations came from top-10 pages in its July 2025 study. I go through the full selection process in how ChatGPT, Perplexity, and Google AI choose which sources to cite.
What Has Google Said About Schema and AI Search?
Google says schema is not required for AI Overviews or AI Mode, but recommends keeping it for rich results and accurate page understanding. Its AI optimization guide lists “overfocusing on structured data” as a mistake, then adds that “it’s a good idea to continue using it as part of your overall SEO strategy, as it helps with being eligible for rich results.”
Google’s position comes down to four statements:
- No special AI schema exists. The AI features documentation says there are “no additional requirements to appear in AI Overviews or AI Mode.”
- Eligibility is about indexing. A page “must be indexed and eligible to be shown in Google Search with a snippet,” per the same page.
- Markup must match the page. The AI features best practices include “making sure your structured data matches the visible text on the page.”
- Business data lives elsewhere too. The AI optimization guide says Merchant Center and Google Business Profile “can help your products and services to be visible in both AI responses and other Google Search results.”
John Mueller summed it up on Reddit as “yes, no, and it depends,” noting that features like shopping need structured data while a comparison site “isn’t going to rank better by adding insurance markup.”
Does Microsoft Copilot Use Schema Markup?
Yes. Microsoft is the only major AI search provider that has publicly confirmed it uses schema markup for its language models. At SMX Munich in March 2025, Bing’s Fabrice Canel confirmed that schema markup helps Microsoft’s LLMs understand content, as reported by Search Engine Land.
Microsoft Advertising repeated the advice in an October 2025 guide, describing schema as code that “helps search engines and AI systems understand your content.” Microsoft’s open NLWeb project also builds on Schema.org data that websites already publish.
If Copilot traffic matters to you, check Bing Webmaster Tools. Its AI Performance report, launched in February 2026, shows how often Copilot and Bing AI summaries cite your pages.
Do AI Chatbots Actually Read the JSON-LD on Your Page?
Mostly not as schema. When ChatGPT, Perplexity, or Claude fetch a page, the tests suggest they read JSON-LD as plain text, if they see it at all, rather than parsing it as structured data. Gemini is the exception in several tests.
The experiments point the same way from different angles:
- searchVIU tested five AI systems on a test product page and concluded (December 2025) that “JSON-LD Schema is NOT read by AI chatbots during direct fetch.” Only Gemini found a JavaScript-rendered price.
- Mark Williams-Cook put a fake company’s address only inside invalid JSON-LD. ChatGPT and Perplexity still returned it, and he concluded (May 2026) they “were not parsing it as schema” but reading the page’s text.
- OtterlyAI asked seven platforms to read its schema directly. Six of seven failed, and only Gemini retrieved it correctly.
There’s also a technical catch. A Vercel and MERJ analysis (December 2024) found “none of the major AI crawlers currently render JavaScript,” including OpenAI’s and Anthropic’s. If a plugin or Google Tag Manager adds your schema with JavaScript, those crawlers never see it. Google can read JavaScript-generated markup, but warns it can make Shopping crawls “less frequent and less reliable.” Put schema in the page’s HTML.
Which Rich Results Has Google Removed?
Google has retired a long list of rich result types since 2023, ending with FAQ rich results on May 7, 2026. Markup for these types still validates, but it no longer changes how your page looks in Google Search.
| Date | What changed |
|---|---|
| August 2023 | HowTo rich results deprecated; FAQ limited to “authoritative, government and health websites” (Seer Interactive) |
| June 2025 | Seven types announced for removal: Book Actions, Course Info, Claim Review, Estimated Salary, Learning Video, Special Announcement, Vehicle Listing (Search Engine Roundtable) |
| September 2025 | Five of those stop showing in results (Google) |
| January 2026 | Practice problem rich results removed (Google) |
| May 7, 2026 | FAQ rich results removed from Google Search (Google) |
This is where a lot of “AI schema” advice goes wrong. Several guides published in 2026 still list FAQ and HowTo as core types for AI. I still add FAQPage markup to articles that have a real FAQ section, because it’s harmless when it matches the visible questions and some systems read page text anyway, but I no longer present it as a ranking or rich result tactic.
Which Schema Types Still Matter in 2026?
The schema types that still matter are the ones that identify who you are, what you sell, and who wrote your content: Organization, LocalBusiness, Person, Article, Product, and BreadcrumbList. They support entity understanding and the rich results Google still shows.
| Priority | Schema type | Why it matters |
|---|---|---|
| Keep | Organization (with sameAs) | Helps Google “disambiguate your organization” and can affect your logo and knowledge panel (Google) |
| Keep | LocalBusiness and subtypes | Name and address are the only required properties (Google); subtypes include LegalService and MedicalClinic |
| Keep | Article with Person author | Google accepts author “url” and “sameAs” to identify the writer (Google) |
| Keep | Product and Offer | ChatGPT shopping uses “structured metadata” such as price and description (OpenAI) |
| Keep | BreadcrumbList | Still a supported Google structured data type and describes your site structure |
| Optional | FAQPage | No rich result since May 2026; fine when it matches visible Q&A |
| Retired | HowTo, Course Info, Estimated Salary, Learning Video, Special Announcement, Vehicle Listing, Book Actions, Claim Review, practice problem | No longer shown in Google Search |
Adoption is lower than most people assume. The 2024 Web Almanac found JSON-LD on 41% of pages, with Organization on 7.16% of mobile pages and LocalBusiness on 3.97%. A clean Organization or LocalBusiness block still puts a small business ahead of most of its competitors.
How Does Schema Help AI Understand Your Business as an Entity?
Schema helps AI systems understand your business as an entity by stating its name, address, people, and official profiles in one consistent place, and by linking those facts with sameAs to the other sites that describe you. That consistency makes it easier for search engines to connect mentions of your brand to the right business.
The useful pieces are:
- Organization or LocalBusiness on your home page, with name, logo, address, phone, and email that exactly match your Google Business Profile.
- sameAs links to your official profiles. Google defines sameAs as “the URL of a page on another website with additional information about your organization,” and accepts more than one.
- Person markup for the founder or authors, linked to the business with worksFor or founder.
- Stable @id values, so your Article author, your Organization, and your Person entries all point to the same entities instead of creating duplicates.
I rebuilt the home page schema on famamun.com this way: a ProfessionalService entry (a LocalBusiness subtype) with my address, phone, email, social profiles, and all eight services, plus a Person entry for me, connected by founder and worksFor. It passed Google’s Rich Results Test with no warnings. It won’t get me cited by ChatGPT by itself, but it removes any doubt about who runs the site and what it offers. Brand mentions across the web do the heavier lifting, which I cover in how LLMs choose sources to cite.
Can Schema Markup Hurt Your Site?
Yes, schema can hurt if it’s misleading or describes content visitors can’t see. Google’s structured data guidelines say “don’t mark up content that is not visible to readers of the page,” and a violation “can result in a manual action.”
The good news is that the damage is limited. Google states that a structured data manual action means a page “loses eligibility for appearance as a rich result; it doesn’t affect how the page ranks in Google web search.” The common mistakes I see on small business sites are:
- Marking up reviews or star ratings that don’t appear on the page.
- Adding FAQPage markup for questions that aren’t visible.
- Two plugins outputting conflicting Organization data, with different names or addresses.
- Schema added only through JavaScript, which AI crawlers can’t see.
Conflicting business details aren’t only a schema problem. Immigration Lawyers USA had three competing Google Business Profile listings for one practice, and consolidating them into one came first, because tidy schema can’t fix an inconsistent presence across the web.
What Should Law Firms, Medical Practices, Local Businesses, and Stores Add First?
Start with the schema that identifies your business, then add the type that matches what you sell. For most small businesses, that’s one LocalBusiness subtype on the home page and accurate Article author markup on blog posts.
- Law firms: LegalService (or Attorney) on the home and location pages, Person markup for each attorney with links to their bar or firm bio pages, and Article author markup on practice-area guides. My law firm SEO services page covers the wider setup.
- Medical practices: MedicalClinic or the closest subtype, Person markup for providers, and author details on health content, since healthcare pages face stricter quality standards.
- Local service businesses: LocalBusiness or the closest subtype with name, address, phone, hours, and sameAs links, matching your Google Business Profile exactly.
- Ecommerce stores: Product and Offer with price and availability in the HTML, plus a Merchant Center feed. ChatGPT shopping also accepts direct product feeds from merchants.
How Do You Measure Whether Schema Made a Difference?
Measure schema the way the better studies do: change a set of pages, leave a similar set alone, and compare both over 60 to 90 days. Without a control group, you can’t tell schema’s effect from normal swings in AI answers, which OtterlyAI saw firsthand when its ChatGPT citations fell 71% during its test for reasons unrelated to schema.
Use three data sources:
- Google Search Console for rich result reports and AI Overview and AI Mode impressions, which I explain in how to optimize for Google AI Overviews.
- Bing Webmaster Tools AI Performance for Copilot citations.
- Monthly prompt checks in ChatGPT, Perplexity, and Gemini, using the routine in how to track whether your content is being cited by AI platforms.
If you want a second pair of eyes on your current schema and AI visibility, my free AI visibility audit checks whether ChatGPT and Google AI Overviews already mention your business. You can also see my AI SEO and AEO services or book a strategy call.
Frequently Asked Questions
Does schema markup help with AI search?
Not directly, based on current tests. An Ahrefs study found adding schema “produced no major uplift in citations on any platform,” though Microsoft says schema helps its Copilot models understand content.
Do I need special schema for Google AI Overviews?
No. Google states there’s “no special schema.org structured data that you need to add” for AI Overviews or AI Mode. Your page needs to be indexed and eligible for a snippet.
Is FAQ schema still worth adding in 2026?
Only where you have a visible FAQ section. Google removed FAQ rich results on May 7, 2026, per its documentation updates log, so the markup no longer changes your search listing.
Does ChatGPT read JSON-LD schema?
Not as structured data, according to independent tests. A searchVIU test found AI chatbots did not read JSON-LD during direct fetch, while ChatGPT shopping does use structured product metadata from feeds.
Which schema type should a small business add first?
Organization or a LocalBusiness subtype on the home page, with sameAs links to your official profiles. Google says Organization markup helps it disambiguate your business in search results.
Can bad schema markup get my site penalized?
It can cost you rich results but not rankings. Google states a structured data manual action removes rich result eligibility and “doesn’t affect how the page ranks in Google web search.”
