SEO »
Schema markup does more than power rich results. Build a knowledge graph, assess entity coverage, and find gaps in AI understanding. #
After presenting a custom schema my team built to evaluate knowledge graphs for university programs, I saw firsthand how differently SEOs view the role of schema markup. Our approach used 23 existing Schema.org entities and more than 60 additional entities to assess gaps in entity coverage.
Schema connects entities (objects, people, concepts, and ideas) to build knowledge graphs that provide deeper semantic understanding. It can also serve as a framework for evaluating a site’s vector embeddings and identifying gaps in entity coverage.
Let’s look at how you can use schema and knowledge graphs to identify and prioritize those gaps.
How knowledge graphs turn entities into context #
A knowledge graph stores entities as nodes and relationships as edges, so machines can understand context and meaning rather than matching keywords. It’s how a system knows that the Tulane Freeman School of Business is an organization that offers business courses taught by a person rather than a string of text.
Enterprise-level businesses use knowledge graphs to remove data silos. In the book “The Knowledge Graph Cookbook: Recipes That Work” by Andreas Blumauer and Helmut Nagy, knowledge graphs are used as the “ultimate linking engine” to create a semantic data fabric for business intelligence.
In this context, your website uses the same recipe that enterprise-level businesses use for internal data. Your website is essentially functioning as a public API through which your brand’s entities, such as organization, location, products, positioning, values, features, and key benefits, are connected to search engines and LLMs.
Dig deeper: When and how to use knowledge graphs and entities for SEO
[ Be the brand AI recommends. See your AI visibility
](https://www.semrush.com/ai-seo/overview?utm_campaign=ic_sel_0101ai&utm_source=searchengineland.com&utm_medium=overlay&onboarding=off) See where your brand appears in AI search, where competitors are winning, and what it takes to become the answer AI recommends.
Treat schema as the on-ramp to the graph #
Schema markup declares entities in a language that search engines and LLMs already speak. JSON-LD explicitly asserts both entities and the relationships between them rather than relying on bots to infer them from website copy.
For example, an executive Master of Business Administration (MBA) degree from New York University should present a coherent body of information rather than a loose collection of unrelated facts. The organization is NYU, which offers a Brand and Digital Strategy course taught by professor and author Scott Galloway, a faculty member in the MBA program at the NYU Stern School of Business.
Stop padding FAQ schema and optimizing for rich results. Website pages are vehicles for nodes and edges that feed the graph. This approach provides critical context about your business, helping search engines and LLMs better connect your products and services with relevant audiences.
Build the graph with markup, vectors, and agents #
Our custom schema is a framework for explicitly asserting all the entities and relationships a prospective student engages with on their enrollment journey. The schema prioritizes these entities by importance.
That declared truth is what we use to analyze the vectorized version of a partner’s website. Vector embeddings of .edu website content are measured against the schema to capture semantic proximity and context that the markup doesn’t state outright.
We use agents to compare the static custom schema markup, semantic proximity, and vector embeddings to ensure comprehensive coverage of all the entities surrounding a university program.
The results surface entity gaps, which can inform content strategy across websites, social media properties, and earned media. The final output is a knowledge graph that highlights both covered and missing entities related to a university program.
We aim to provide robust context that helps shape and frame the value of our programs through a search or conversation. This approach is directly aligned with Google’s patents on understanding entities.
Google’s need to understand context is driven by a desire to deliver a more relevant answer to its end user. Building your knowledge graph adds context that can help search engines and AI systems better understand your business.
*Dig deeper: *Google’s LLM patent suggests a new goal for SEO: Teaching AI who you are
Get honest about schema and AI visibility #
In March 2025 at SMX Munich, Fabrice Canel, principal product manager at Microsoft Bing, confirmed that Copilot uses schema markup to understand content.
When we consider how other LLMs use search indexes for grounding, schema markup becomes an important factor in connecting entities back to your site. Semrush has also published guides showing correlations among cited pages.
Other research points in a different direction. A 2025 study by Search Atlas found that schema doesn’t affect LLM visibility, while a study by Mark William-Cook showed that LLMs don’t read on-page schema. However, those studies look at performance and citations rather than entity connections.
These findings may seem conflicting, but they point to a distinction between using JSON-LD as infrastructure for machine understanding and treating it as a GEO hack for increasing citations. JSON-LD provides explicit information about entities and relationships that can help machines understand your site. Visibility is a byproduct of being understood.
Use schema and vector embeddings to find entity gaps #
The custom schema I discussed above is specifically designed for higher education, but you can build your own. Use Schema.org as your library to build a robust collection of entities that matter to your product, service, or business.
Does Schema.org map to all your important entities? In our case, Schema.org covered only 23 higher education-specific entities, leaving significant gaps. We filled the gaps by creating our own entities.
Start with the ideal entity model. In a perfect world, what does that complete set of information look like? This exercise will help you determine what can be built using custom JSON-LD in addition to established entities.
With that custom schema, you can compare it against vector embeddings to understand entity coverage. Prioritize your entity gaps by focusing on what drives the most value.
Dig deeper: How schema markup fits into AI search — without the hype
Measure entity visibility, not just rich results #
As you develop a content strategy for entity coverage, it’s important to use prompt-tracking and brand sentiment tools to monitor AI visibility. Monitor your priority entities to understand how you appear across different models.
Brand sentiment tracking is especially important because it allows you to see shifts among your covered entities. Is your content matching how people are actually talking about the entity?
AI visibility and brand perception aren’t complete metrics on their own. As you see shifts in visibility and perception, compare them with conversions and performance. As citations grow, does the volume or quality of leads rise?
[
If AI can’t find you, customers won’t either.
See your AI visibility
](https://www.semrush.com/ai-seo/overview?utm_campaign=ic_sel_0102ai&utm_source=searchengineland.com&utm_medium=overlay&onboarding=off) Track your visibility across AI search, uncover missed opportunities, and grow your presence where customers are asking questions.
Schema as infrastructure for AI understanding #
Schema markup isn’t an overnight solution for driving AI citations. It’s infrastructure for helping search engines and LLMs understand your site by declaring entities and relationships, building your knowledge graph, and uncovering gaps in entity coverage.
Contributing authors are invited to create content for Search Engine Land and are chosen for their expertise and contribution to the search community. Our contributors work under the oversight of the editorial staff and contributions are checked for quality and relevance to our readers. Search Engine Land is owned by Semrush. Contributor was not asked to make any direct or indirect mentions of Semrush. The opinions they express are their own.