There’s been some online debate about whether Schema Markup actually helps brands appear in AI-generated answers.
Some recent studies have found little or no correlation between the presence of Schema Markup and AI citations. From there, people have quickly concluded that brands should stop investing in Schema Markup for AI visibility altogether.
But that conclusion overlooks the important distinction that not all Schema Markup is created equal.
There is a significant (and strategic) difference between adding basic Schema Markup to a webpage and using advanced, connected Schema Markup to define your organization’s entities and relationships in a Content Knowledge Graph that lives on your website.
If we evaluate all Schema Markup as though it provides the same semantic value, we risk implementing and measuring the wrong thing.
Basic Schema Markup Was Built for a Different Job
For years, Schema Markup strategies have largely focused on helping search engines qualify content for rich results.
A recipe gets Recipe markup, a product gets Product markup, etc. This type of Schema Markup is still valuable, but simply having Schema Markup on a page does not necessarily mean you have clearly defined what that content means or how the entities within it relate to the rest of your organization.
AI systems have a much bigger interpretation problem to solve. They need to understand context-heavy questions like:
- Who is this organization?
- What products or services does it offer?
- Which locations provide which services?
- Who are its experts?
- How are these people, products, services, locations, and topics related?
- Which information should be trusted when sources conflict?
Answering these questions requires advanced, semantic Schema Markup that defines entities and connects their relationships across your website.
The Difference Is a Content Knowledge Graph
Basic Schema Markup describes what exists on a page. Advanced Schema Markup connects those descriptions across your website to create a Content Knowledge Graph.
With basic Schema Markup, a healthcare organization, for example, might identify a page as being about a physician, procedure, specialty, or hospital location. This gives machines structured information about that page, but often stops short of explaining how it relates to the rest of the organization.
Advanced Schema Markup goes further. It defines the key entities across your website and explicitly connects the relationships between them, transforming disconnected structured data into a Content Knowledge Graph.
For example, looking at that healthcare organization again, they can define a physician in their Schema Markup, connect that physician to their medical specialty, connect the specialty to a procedure, and connect that procedure to the locations where it is offered.
These connections are strengthened through Entity Linking, which helps machines disambiguate what an entity means by connecting it to other defined entities across your website (Internal Entity Linking) and to known entities in authoritative knowledge bases across the web such as Wikidata (External Entity Linking).
The result is a connected semantic data layer that explicitly tells machines what your organization knows, what those things mean, and how they relate.
That context is what makes advanced Schema Markup particularly valuable for AI. Instead of requiring an AI system to infer meaning and relationships from disconnected content, a Content Knowledge Graph makes those relationships explicit and machine-readable, creating a stronger foundation for accurate AI understanding, visibility, retrieval, and AI-powered experiences.
Our Data Shows Advanced Schema Markup Supports AI Understanding
On Schema App’s own website, we measured a 19.72% increase in AI Overview visibility after implementing robust Entity Linking around a prioritized group of entities.
We also saw growth in the number of relevant keywords for which our content appeared in AI Overviews and stronger alignment between those citations and the topics where we wanted to establish authority.
More recently, as we continued to strengthen Entity Linking and the connections within our Content Knowledge Graph, our AI Overview visibility increased another 32% quarter over quarter.
We are seeing similar patterns with enterprise customers.
A Fortune 500 Schema App customer increased its average market share of tracked keywords receiving AI Overview citations from 27.5% to 36% after implementing Schema App with linked entities. The customer confirmed that no other major SEO initiatives were underway during the period.
MasterControl has also observed that pages supported by Schema Markup are consistently the pages being cited by AI systems. After implementing Schema App, the company saw a 29% increase in domain-wide search visibility, while key blog page sets using Entity Linking experienced a 39% increase in impressions.
These examples do not prove that adding any Schema Markup to any webpage will automatically earn an AI citation. They show why that may be the wrong hypothesis to test in the first place.
Schema Markup Can Also Influence the Accuracy of AI Responses
Visibility is only one part of the equation. For enterprises, being cited by AI is not particularly valuable if the information AI provides is wrong.
Wells Fargo experienced this firsthand when a Google AI Overview incorrectly stated that one of its branches had permanently closed. The AI Overview was relying on a decades-old third-party article instead of Wells Fargo’s current branch information.
After we deployed Schema Markup to the relevant branch locator pages, the incorrect result resolved within minutes. Google stopped citing the outdated article and began referencing Wells Fargo’s accurate, marked-up branch locator pages instead.
When it comes to leveraging Schema Markup for AI, the overarching goal has always been to “get cited,” but it should also be to give machines an accurate, explicit, machine-readable source of information about your organization.
Schema Markup is Infrastructure for AI
The value of advanced Schema Markup also extends beyond Google AI Overviews.
When Schema Markup is used to build a Content Knowledge Graph, the resulting semantic data layer becomes an asset that organizations can activate and reuse across other AI applications.
For example, a Content Knowledge Graph can be made available to AI agents and copilots through technologies such as Model Context Protocol (MCP), giving those systems access to structured, context-rich organizational knowledge.
The same principle applies to AI experiences organizations build themselves.
An internal chatbot or conversational search experience is only as reliable as the information it can retrieve and understand. Giving these systems access to a governed, brand-controlled semantic data layer provides clearer context about your organization than asking an LLM to continuously infer relationships from disconnected webpages.
This is why the conversation needs to move beyond asking whether Schema Markup works for AI, and toward “What calibre of Schema Markup are you using, and how are you measuring its value?”
Schema Markup Has Evolved; Your Strategy Should Too
We want to be clear that basic Schema Markup still has a role to play in search. But applying a few Schema.org properties to individual pages and expecting an immediate increase in AI citations misses the larger opportunity.
Advanced Schema Markup can help organizations define their entities, connect their content, govern how their information is represented, and create a machine-readable semantic data layer.
That foundation can support visibility in AI search, help AI systems represent your organization more accurately, and provide trusted organizational knowledge to the AI experiences and agents enterprises are building themselves.
So, should you stop doing Schema Markup for AI?
If you’re treating Schema Markup as a collection of disconnected tags designed only to earn search features, it may be time to rethink the investment.
But if you’re using Schema Markup to build a connected, governed representation of your organization that machines can understand and reuse, that is more than Schema Markup. That is building a semantic foundation for AI.
Andrea Badder is the Content and SEO Manager at Schema App. She creates educational content that helps marketing teams understand how structured data and Content Knowledge Graphs drive visibility and accuracy in search and AI. Prior to joining Schema App, Andrea worked as a brand strategist and copywriter at a marketing agency. She is a graduate of the University of Guelph.