Historically, Schema Markup has been used to target rich results in SERPs. However, we have recently seen a shift in the SERP landscape, with AI overviews now taking up valuable real estate at the top of SERPs and lowering the priority of rich results. The question now is – how can we optimise content to be best understood by AI, so this can be featured not only in the AI overview space but also in AI search engines?

Schema is a machine-readable markup language that can be added to a page’s HTML and is used to define what is on a page for better comprehension by search engines [5]. There are various types of Schema (e.g. Product, Organization, ImageObject) – with a full list available on schema.org/ – but it is important to note only some have associated rich results, and that these are not guaranteed even with schema markup added to your page. However, don’t let this put you off from adding other types of schema to your site – the more information you can markup about your content the better!

Schema can be added directly to page HTML, or often there are options to add this directly through your CMS. Plugins such as Yoast can also be used to generate Schema – this code can often be taken as a starting point to create the more interconnected graphs we will touch on later in this article.

When writing Schema, we recommend making good use of tools like Google’s Structured Data Markup Helper or the JSON-LD Playground to validate code as you write it. Google’s Rich Result Validator will show if your Schema markup (either through a page URL or code snippet) is correct and eligible for rich results, whilst Schema.org’s Validator will show if a page’s Schema implementation or code snippet is correct and error-free.

Schema in the context of Semantic (Entity) SEO

To understand how to optimise for AI search, we must first understand the shift from lexical to semantic search algorithms. Lexical algorithms rely solely on keyword matching, whilst semantic algorithms focus on comprehending natural language, alongside the meaning and intent behind a query, to give answers that go beyond surface-level associations [2]. Semantic search aims to deliver more relevant, helpful, and tailored results for users, resulting in a better user experience and more intuitive search behaviours.

Semantic search structures and understands content through modelling this into entities, their properties, and the relationships between them [3]. Here, an entity refers to a specific concept in the real world – a person, place, organisation, idea, etc. Google combines all these components together into a knowledge graph – a graphical representation of entities, their attributes and the links between them.

AI platforms such as ChatGPT and Gemini rely on semantic search and schema markup to interpret and process information, as this allows them to extract information and comprehend content much faster and with fewer computational resources [6]. Therefore, optimising your Schema markup for semantic search will also help optimise for AI.

Essentially, accurate and thorough Schema markup presents semantic value to search engines – it adds greater contextual meaning and defines relationships between entities on a site [1]. Adding Schema markup provides this information to search engines in a quick, clear and easily understood way, meaning search engines do not have to process and infer this information themselves.

Historically, Schema has not been used as extensively as it could to define the relationships between entities, as adding independent Schema types was enough to go after rich results. However, when looking to optimise for semantic search, the focus has shifted less from defining entities themselves to defining the relationships between them, focusing on building up Google’s knowledge graph for your content to ensure this is properly and thoroughly understood.

In real terms – we want our Schema Markup to be one connected graph, rather than a series of separate blocks of code so search engines (and AI search!) can best understand these relationships [4].

For example, you may have:

• NewsArticle

• WebPage

• WebSite
• Organization

As 4 separate blocks of code on your site, this may be sufficient to try and gain a rich result, but this does not define the relationships between any of these entities.

What we would instead want is:

• NewsArticle

• • Is MainEntity of WebPage
    • Belongs to Website
      • Owned by Organization 

We (and search engines) can now clearly see how these entities are linked in one graph, helping convey greater meaning and optimising for semantic search.

When looking to future-proof your SEO strategy, you need to ensure that search engines can truly comprehend your content, not just parse it and identify keywords, and Schema Markup is an important tool for this. We have seen rich result types come and go, but Schema as a tool is important to future-proof your SEO strategy in 2025 and beyond – laying the fundamentals of getting your content fully and accurately understood by semantic search models and AI search engines.

Interested in Schema markup for your site? Contact Varn today to talk to our team of experts and discuss your AI search strategy!

References:

[1] https://www.schemaapp.com/schema-markup/evolving-role-of-schema-markup/

[2] https://myscale.com/blog/semantic-search-vs-lexical-search-key-differences/

[3] https://www.wix.com/seo/learn/resource/semantic-seo

[4] https://yoast.com/why-schema-needs-to-be-a-graph/

[5] https://www.wix.com/seo/learn/resource/structured-data-for-seo

[6] https://www.npgroup.net/blog/role-of-schema-markup-in-ai-friendly-websites/