With AI Overviews and ChatGPT answering questions directly, does schema markup still matter? Short answer: yes, but the reason changed. Schema markup for AI search is no longer just about rich results; it’s about making your content machine-readable for the AI systems that now decide what gets cited.
A year ago, schema markup was about winning rich results: star ratings, FAQs, and event details in Google’s search results. Then AI Overviews arrived, ChatGPT Search launched, and suddenly the question I hear from clients is: “If AI just summarizes everything, why bother with structured data?” Fair question. Here’s the honest answer.
What Actually Changed
Traditional search showed ten blue links; schema helped you stand out among them. AI search synthesizes answers from multiple sources and cites a few. The game shifted from standing out in a list to being selected as a source. And selection is exactly where structured data helps: it tells machines precisely what your content is, who wrote it, and what facts it contains. No interpretation required.
Think of it this way: AI systems are drowning in unstructured web content. Schema markup is you handing them a labeled, organized file instead of a messy pile of paper. Which one gets cited? For the full background on AI search, see how to optimize your WordPress site for AI search.
3 Reasons Schema Still Matters (Updated for AI Search)
1. It feeds the knowledge graph AI answers come from
AI Overviews don’t invent facts. They retrieve them. Google’s knowledge graph, built substantially from structured data across the web, is one of the sources those answers draw on. Organization, LocalBusiness, Product, and Person schema feed entity understanding: this business exists, it’s located here, these are its hours, these people wrote this content. Without structured data, you’re asking the AI to figure all that out from prose.
2. E-E-A-T signals are machine-readable through schema
AI systems selecting sources look for credibility signals: who authored this, what are their credentials, when was it published, is it still current? Author and Article schema with author, datePublished, and dateModified properties make those signals explicit. A well-marked article by a named expert with a recent modified date is easier for an AI to trust than an anonymous undated page, because the trust signals are structured, not buried in paragraphs.
3. Rich results aren’t dead, they’re just not the whole story
Reports of rich results’ death are exaggerated. Product rich results, local business panels, event listings, and recipe cards still appear constantly in traditional search, which still handles the majority of queries. Schema remains the price of admission for those. The AI era added a reason for structured data; it didn’t remove the old one.
What to Prioritize Now
If you’re adding or auditing schema in the AI era, prioritize in this order:
- Organization + LocalBusiness: entity identity: who you are, where you operate.
- Article/BlogPosting with author and dates: credibility and freshness signals for AI source selection.
- FAQPage: question-answer pairs are the native format of AI answers. Well-structured FAQs map directly onto the questions AI systems answer.
- Product/Offer: if you sell anything: pricing and availability in structured form.
- SpeakableSpecification: marks sections suitable for voice/AI reading. Early days, but directionally right.
The practical how-to, including Rank Math setup, is in my schema markup guide for WordPress. The strategy around it (direct answers, quotable stats, llms.txt) is in the AI search optimization guide.
What Schema Can’t Do
Honest limits: schema won’t make thin content rank, won’t force an AI to cite you, and won’t compensate for content nobody wants to read. It’s a clarity layer on top of genuinely good content: it makes great content easier for machines to understand and trust. The content still has to deserve the citation.
The Bottom Line
Schema markup in the AI era isn’t about decorating search results. It’s about speaking the machines’ language at the exact moment machines decide what to cite. It takes an hour to set up properly on WordPress and it compounds across both traditional and AI search. That’s about the highest ROI hour in SEO right now.
Want your site’s structured data audited and fixed? It’s part of my SEO services. Get in touch.
Frequently Asked Questions (FAQ)
Does schema markup help with ChatGPT citations?
Indirectly but genuinely. ChatGPT’s search features retrieve from web sources, and structured data makes your content’s facts, authorship, and freshness unambiguous, which helps retrieval and trust assessment. It’s one signal among many, not a magic citation button.
Is FAQ schema still worth adding after Google limited FAQ rich results?
Yes, arguably more than before. FAQ rich results in traditional search are limited, but FAQ-structured content maps perfectly onto AI answers, which are fundamentally question-answering. The format outlived the rich result.
What’s the easiest way to add schema on WordPress?
Rank Math or similar SEO plugins generate most schema types from settings pages. No code needed. See the step-by-step in my schema markup guide.
