Is Schema still relevant in 2024?

Wonderful debate here. Let’s set the record straight—if you’re still stuck thinking schema markup is just a checkbox for rich results, you’re not only missing the point but also losing the game. The real problem isn’t schema—it’s how people think about schema. If your entire strategy is to slap some structured data on your pages and pray for a featured snippet, congratulations: you’re playing the wrong game.

Schema markup is not dead—it’s evolving. If you’re not using schema as the foundation for Semantic SEO and a broader data strategy, you’re already falling behind. LLMs—yes are trained on both structured and unstructured data. Guess what? Schema matters here too. It’s not just about helping crawlers but about creating and refining entities that power modern AI systems.

Let’s talk use cases:

1. Brand-Entity Optimization

Think you can work without schema to boost your brand-entity? Sure, you can—but good luck connecting your brand’s data across the web. If you want to win, you’ll invest in building semantic data: train your models, create content that answers real questions, and ensure your brand’s entities are accessible, crawlable, indexable and linked. If your data isn’t where AI models like Google, Meta AI or Bing are looking, you’ve already lost. This is about strategy. Stop minimizing investments and start maximizing returns. Are you still using Google? Most of us are. Take a closer look at the carousels your company is featured in and analyze the facets Google is presenting to searchers. Are you chatting with ChatGPT? What’s the share of voice of your entities when your clients engage with it?

Spend time analyzing this. We’re doing it—using the same entities we define in the JSON-LD behind our websites. This isn’t about schema; it’s about entities.

LLMs can be steered with monosemantic features. Do you want to interpret how a language model functions? Do you want to know if the neurons in an LLM are aligned with your messaging? You’ll need entities. I’ll share more on this very soon.

2. Information-Rich Environments

Travel? E-commerce? You’re swimming in data, but can AI make sense of it? Without semantic structure, probably not. Schema isn’t just markup—it’s your competitive edge. Platforms like Google Maps Content Partners, Merchant Center, Data Commons, and Transit thrive on high-quality, structured data. Feed these ecosystems, and you unlock unparalleled opportunities for visibility and engagement. Semantic value, not just structured data, is what AIs are hungry for. Feed it or fade away. Is schema markup enough? No, but it’s the beginning of the journey.

3. AI-Driven Engagement

Building conversational experiences? Recommending related content? Turning your website into embeddings isn’t enough. You need semantic data to extend schema.org with facets that matter to your audience. For example:

• “What are the upcoming events in Zurich where I can wear traditional lederhosen?” This query requires representing facts as discrete triples to identify the event’s location, date, and cultural context, while embeddings capture the nuanced relationship between “events,” “Zurich,” and “lederhosen.”

• “Find me a red wine from Tuscany aged in oak barrels that pairs well with beef and costs under 20 euros.” To answer this, you need triples for wine properties like origin, aging process, and price, while embeddings handle the complex relationship between “red wine,” “beef pairing,” and the intent behind the budget constraint.

• “What products are available at local stores near me that offer eco-friendly certifications?” This combines triples to define product attributes, certification types, and store locations with embeddings to contextualize “eco-friendly” preferences in a search query.

If you’re not building this level of semantic precision, someone else is—and they’re taking your users with them.

Combining discrete knowledge representation (triples) with continuous embeddings isn’t just a nice-to-have; it’s the backbone of delivering rich, meaningful user experiences in modern AI-driven search and conversational systems.

4. Scaling Content Responsibly

Scaling content generation without spamming the internet is the future. Leaders like Microsoft are pushing Graph RAG (Retrieval-Augmented Generation grounded in graphs), while Google is evolving RAG into RIG (Retrieval Interleaved Generation) to ground responses in structured, reliable statistical data. Both are heavily investing in graph-based systems and schema and structured data are the backbone of these innovations. If you’re ignoring schema, you’re ignoring what powers these advancements.

So here’s the reality: schema markup isn’t “over.” It’s just not what most people think it is.

This isn’t about gaming Google; it’s about building smarter systems that understand and connect your data. If you’re still obsessing over featured snippets, fine, keep doing that (it does work)—but the market is moving on.

The winners are the ones building data for humans and AI alike. If you’re not investing in evolving schema, you’re falling behind—and fast.

It’s time to rethink your approach. Start by analyzing how your data connects with the systems shaping modern search and AI.

Are you building for the future? If not, now is the moment to align your strategy with where the market is headed.