The Future of Search: AI, Machine Learning & Schema Markup

Over the past few months, there’s been a lot of buzz around ChatGPT, the “New Bing” and Google Bard. These new innovations in search are powered by machine learning and artificial intelligence (AI).

These topics were also discussed at length during the keynote presentations at Pubcon Austin 2023. Our CEO, Martha van Berkel was there to hear it and left Pubcon with an insight or two on the future of Schema Markup and AI-powered search.

The relationship between Schema Markup & AI 

This past week, Martha presented at Pubcon Austin 2023 on the Top Ways to Use Schema Markup. During the presentation, she discussed how Schema Markup relates to Artificial Intelligence (AI) and the importance of connected Schema Markup in an era of AI-powered search.

Connected Schema Markup helps build your Knowledge Graph

When you implement connected Schema Markup across your site, you are essentially building a knowledge graph.

“A knowledge graph is defined as a collection of relationships between things  defined using a standardized vocabulary, from which new knowledge can be gained through inferencing.”

Schema.org is the standard vocabulary used by search engines to understand the content on a page and when you use it to define relationships between entities on a website you are creating a knowledge graph for that organization. In fact, when you look closely at Schema.org, you can see how it was build with creating knowledge graphs in mind.

When you look at the Schema.org vocabulary definitions, you’ll often see many properties expecting the value of the property to be another Type instead of a text or URL value.

Using Physician as an example, the Expected Type for the hospitalAffliation property is another schema.org Type called Hospital. By embedding the Hospital Type under the hospitalAffliation property, you are defining the relationship between the Physician and the Hospital that they are affiliated with.

This example showcases how the Schema.org vocabulary defines relationships between things on your website. When you connect and define these relationships using Schema Markup, you are building your knowledge graph.

At Schema App, we know the importance of creating connected Schema Markup and building a reusable knowledge graph. Our enterprise solution utilizes semantic technology that creates and stores your Schema Markup and web content in a knowledge graph.

But what do knowledge graphs have to do with AI and Machine Learning? Why is it important for organizations to develop their knowledge graph using Schema Markup?

We asked ChatGPT why knowledge graphs are important and here’s what it had to say.

“Knowledge graphs are a powerful tool for representing and organizing information in a structured way.

Here are three reasons why knowledge graphs are important:

1. Integration and Interoperability: Knowledge graphs allow for the integration of heterogeneous data from different sources, formats, and languages. This enables organizations to gain a holistic view of their data, making it easier to identify patterns and insights that might not be visible when data is siloed. Knowledge graphs also enable interoperability between different systems, making it easier to share and reuse data across different applications.
2. Semantic Search and Analysis: Knowledge graphs use semantic relationships to represent information, making it easier to perform advanced search and analysis tasks. For example, knowledge graphs can be used to perform complex queries that take into account the relationships between different entities, such as “find all the people who have worked at companies in the same industry as Microsoft.” This can help organizations identify patterns and insights that would be difficult to discover using traditional search and analysis methods.
3. Machine Learning and AI: Knowledge graphs are a foundational technology for machine learning and AI. By representing information in a structured and standardized way, knowledge graphs make it easier for algorithms to extract insights and make predictions. For example, knowledge graphs can be used to train natural language processing ****** (NLPs), or to build recommendation engines that suggest products or content based on a user’s preferences and behaviours.”

(Source: ChatGPT, Open Ai)

Search engines leverage your Schema Markup and knowledge graph as data sources to train their machines and infer new knowledge. By developing your organization’s knowledge graph, you can prime your organization’s web data to be ‘AI-ready’.

Earlier this year, Ryan Levering, Google’s champion for structured data, said the following with regard to what Google wants from Schema Markup.

Also, over time richer/correct semantics will favour more connected graphs.
– Ryan Levering, Google (Source: Mastodon)

Even though Google has yet to release any official documentation around connected Schema Markup, Levering’s comment indicates its growing importance in the world of search.

Our sentiments on connected Schema Markup were also echoed by Fabrice Canel, Principal Program Manager for Bing, at his keynote presentation in Pubcon Austin 2023.

SEO recommendations for Bing AI Search

During his keynote presentation, Canel offered valuable tips and insights on optimizing for Bing’s new AI search engine. Even though AI search is in its infancy days, Canel shared that one of the ways SEOs can prepare for this new AI-enabled search is by writing great content and annotating with Schema Markup. 

In a later slide, he further elaborated on what they mean by great content and Schema Markup. They specifically mentioned using ‘Semantic markup’ to convey information about the pages. 

Semantic markup is also known as connected Schema Markup, where you define the relationships between the content on your pages and other definitions on the web using the properties defined in Schema.org.

This goes to show that connected Schema Markup is important for AI search engines and SEOs need to invest in it. It is also why our team at Schema App constantly emphasizes it when building a Schema Markup strategy for our customers.

How AI will transform the search experience

During his keynote presentation, Canel also shared about the various types of search queries and how the search engine results will vary to best satisfy the user’s query.

Our takeaway from it is that the new Bing Chatbot experience will suit some queries, while others other queries will be better answered with a table or a version of today’s search results.

For example, for queries such as ‘Tell me all the hotels in the Dominican that have waterslides’, users might be satisfied with a chatbot summary answer and even accept a margin of error.

On the other hand, for queries such as ‘What is the recovery time from a hip surgery”, users might want to read different articles on the subject and personally determine who the subject matter expert is before accepting the answer.

Over time it will be interesting to see how these different experiences and types of searches evolve with this new AI chat technology.

Despite the recent buzz, AI and machine learning are not new to search. Gary Ilyes from Google kicked off Pubcon with his keynote presentation on how AI dates back to the bronze age, how these concepts are already deeply entrenched in how we conduct business, and how they will continue to evolve.

We really enjoyed seeing how existing industries today are using machine learning and AI through the tools and process automation that are already adopted.

However, Ilyes did not comment on Bard or whether Google would be releasing a Chatbot in response to the New Bing so we’ll just have to wait and see.

Schema App & AI

At Schema App, we also utilize AI and machine learning in our tools. We use it for our Linked Entity Recognition feature and our Schema Performance Analytics tool.

Because of our passion for semantic technology, your data is stored in a knowledge graph when you create your Schema Markup with Schema App. We then layer on additional AI capabilities to help you add more meaning to your content.

For example, Schema App’s Linked Entity Recognition, allows our technology to create connected Schema Markup using Natural Language Processing to connect your content to known entities in Google’s Knowledge Graph and Wikidata. This provides context to the content, connecting content using the sameAs link, or more flexibly with mentions, about, category, etc.

An upcoming release will also include a Medical BERT conceptual model that Healthcare companies can use to advertise all their specialties. We’re also working on a feature for Schema Performance Analytics to generate AI insights from the performance data, and will be releasing it in Beta shortly.

Start preparing for an AI-powered search experience

AI and Machine learning are here to stay and will continue to gain prominence in the search experience. Thankfully, the evolution within search will not happen overnight. It will likely evolve over the next few years.

However, organizations need to ready themselves for what’s to come. As you adopt, deploy and manage your Schema Markup to achieve a rich result, you also want to ensure that you’re doing it semantically to build a connected knowledge graph. That way, you can lay the foundations to be relevant for search engines and perform well in this new search experience.

As a semantic technology company, Schema App is excited to provide you with the expertise and tools to do this in a scaleable, manageable way with measurable results. If you need help creating connected Schema Markup, get in touch with us today to find out how we can help you prepare for this new search experience.

Martha is the CEO and co-founder of Schema App. Schema App is an end-to-end Schema Markup solution that helps enterprise SEO teams create, deploy and manage Schema Markup to stand out in search. She is an active member of the search engine optimization community, and the work that she does through Schema App is helping brands from all over the world improve their organic search performance.