Why JSON-LD Schema Matters for Getting Your Website Cited by AI

JSON-LD helps machines read who you are and what you offer, and Microsoft says schema markup helps its language models understand content for Copilot. What no one has shown is that adding it raises how often AI systems cite you, and Google says AI Overviews and AI Mode need no special schema at all. The bottom line: treat JSON-LD as accuracy infrastructure for your business facts, fix it because wrong or missing facts hurt you, and do not expect it to work as a citation lever on its own.

For a full definition, read our answer engine optimization guide.

0special schema types Google says AI Overviews and AI Mode require

Mar 2025Microsoft's Fabrice Canel said schema helps Copilot's language models understand content

1,885pages in Ahrefs' controlled test that added JSON-LD, with no meaningful citation lift

16.7% to 54.2%GPT-4 accuracy on enterprise questions, raw SQL versus a knowledge graph, in a data.world benchmark

Where JSON-LD comes from

Google, Microsoft, Yahoo, and Yandex launched the shared schema.org vocabulary in 2011 so that a page could describe things, a business, a product, an article, in terms every search engine understood. Google had already announced its Knowledge Graph in May 2012 with the line "things, not strings," which is the idea behind all of this: engines wanted to know that a page is about a specific entity, not just contains a word. JSON-LD, a way of writing that vocabulary as a self contained block of JSON, became a W3C Recommendation in January 2014, and Google now recommends it over the older Microdata and RDFa formats because it lives in one place instead of being woven through the visible HTML.

What each platform has actually said

Platform	Public position on schema	Source
Google (AI Overviews, AI Mode)	A page needs to be indexed and snippet eligible. No additional technical requirements, no special schema.org markup, no machine readable or AI text files needed. Structured data should match visible text.	Google Search Central, AI features and your website
Microsoft (Bing, Copilot)	Fabrice Canel, a Bing principal product manager, said at SMX Munich in March 2025 that schema markup helps Microsoft's language models understand web content.	Reported by Machina and others
OpenAI, Anthropic, Perplexity	One review of their public documentation found no published guidance on how their retrieval systems treat JSON-LD.	Passionfruit review

What the evidence shows

Three findings matter, and none of them says what most schema marketing implies. A December 2024 Search/Atlas study found no correlation between schema markup coverage and citation rates across AI platforms. Ahrefs ran a controlled test comparing 1,885 pages that added JSON-LD with 4,000 control pages and found no meaningful increase in citations in AI Overviews, AI Mode, or ChatGPT, though every page in that test was already cited more than 100 times, so it cannot say whether schema helps a page get noticed in the first place. The most encouraging result comes from outside SEO: a data.world benchmark by Sequeda, Allemang, and Jacob had GPT-4 answer enterprise questions two ways. Against a raw SQL database it scored 16.7 percent accuracy, and grounded in a knowledge graph of the same data it scored 54.2 percent. That is evidence that structure improves machine accuracy when the machine can use it, not evidence that a schema block on a web page earns a citation.

The useful distinction is comprehension versus citation. Schema plausibly helps a system get your facts right, such as your name, address, hours, and services. Whether a system then chooses to cite you depends on retrieval, relevance, and corroboration, which markup does not control.

What to implement

Type	Why it earns its place
Organization or LocalBusiness (most specific subtype)	States your name, address, phone, hours, service area, and sameAs links to your other profiles, which is how a machine ties your site to your business.
Article or BlogPosting	Attaches headline, author, and dates to each post, so a system can judge who wrote it and how current it is.
BreadcrumbList	Cheap to add and still shown in results.
Product, Service, or SoftwareApplication	Describes what you sell in fields a machine can read, when it matches the visible page.

Only mark up what is visible on the page. Google's guidance is explicit that structured data should match the text a visitor can read, and markup describing content that is not there is a liability, not an advantage.

How our own scanner treats it

Our scan gives schema the largest weight of its eight checks, 22 of 100 points, and passes a page that has JSON-LD with a business entity type plus at least one supporting type. We weight it highest because it is the item we can verify on the page and fix, not because a study shows it moves citations. That is the honest framing for any readiness score: it lists what you control, and only real testing of customer questions shows what changes what customers see.

Questions people ask

Does JSON-LD help you get cited by ChatGPT or Google AI Overviews?

There is no published evidence that adding it increases citations. Google says no special schema is required for AI Overviews or AI Mode, and Ahrefs' controlled test found no meaningful lift, though that test only covered pages already heavily cited.

Does Microsoft use schema markup for Copilot?

Microsoft's Fabrice Canel said in March 2025 that schema markup helps Microsoft's language models understand web content for Copilot.

Which schema types are worth adding?

Organization or the most specific LocalBusiness type, Article on blog posts, BreadcrumbList, and Product, Service, or SoftwareApplication where they match the visible page.

Can wrong schema hurt?

Yes. Markup that does not match visible content can create parity problems and give machines wrong facts, so keep it accurate and current.