Does Schema Markup Increase AI Citations

Does schema markup actually help with AI citations?

You’ve probably heard that adding schema markup to your website improves your chances of being cited by AI tools like ChatGPT and Google AI Overviews. New research from Ahrefs challenges this assumption with data that might surprise you.

Schema markup AI citations showed no improvement in controlled testing

Ahrefs tested whether adding schema markup to pages already cited by AI would boost their citation rates. The results were clear: it didn’t work.

The study tracked 1,885 web pages that added JSON-LD schema markup. Each page was matched against control pages that never added schema. Citation changes were measured across Google AI Overviews, Google AI Mode, and ChatGPT.

No platform showed a meaningful citation increase after schema was added. In fact, Google AI Overviews showed a small decline of 4.6% relative to the control pages.

Why previous assumptions about schema markup AI citations were wrong

Here’s the confusing part: pages cited by AI are roughly three times more likely to include JSON-LD schema. This correlation made many SEO professionals assume that schema directly improves AI visibility.

Ahrefs dug deeper. They analyzed 6 million URLs and found that sites with schema markup tend to invest more in content quality and earn more backlinks. The correlation exists, but it’s not causal.

To isolate schema’s true effect, they ran a controlled comparison. Each page that added schema was matched with three control pages from different domains. These control pages had similar citation levels but never added JSON-LD.

Using their Brand Radar tool and Agent A, Ahrefs conducted a matched difference-in-differences analysis. They measured citation changes 30 days before and after schema addition. The results were consistent across four different testing methods.

While Ahrefs used their Brand Radar tool for this analysis, publishers and SEOs looking to track their own AI citation performance can use similar monitoring solutions to establish baselines before making optimization changes. Tools like AI Mentions allow users to track when their brand, website, or content appears in AI responses across multiple platforms. This type of real-time monitoring enables testing optimization strategies against their own baseline data.

The Google AI Overview decline needs context

The 4.6% decline in Google AI Overviews deserves explanation. Both treated pages (with schema) and control pages were already declining before schema was added. The treated pages declined slightly faster, but the difference is small.

Most pages in the sample received hundreds of daily citations. The decline represents about 12 fewer citations per page per day. Ahrefs notes this could reflect a small negative effect from schema, or it could be coincidence.

The report doesn’t draw a firm conclusion either way. The decline is statistically notable but practically small.

What the schema markup AI citations study couldn’t test

Every page in the dataset already had 100+ AI Overview citations before schema was added. These pages were already visible to AI systems, being crawled and surfaced regularly.

This creates an important limitation. The study can’t tell us whether schema helps pages that aren’t yet visible to AI. For pages not currently cited, schema might still help with crawling, parsing, or indexing.

The report acknowledges other limitations. Pages adding JSON-LD often change other elements at the same time. This makes it hard to separate schema effects from other improvements. All schema types were pooled together, so some specific types might perform differently than others.

The 30-day measurement window might miss slower effects that take longer to appear.

AI systems ignore schema when fetching content

A separate experiment cited in the report tested whether five AI systems used schema markup when fetching pages in real time. The results were telling: none of them did.

The AI systems only extracted visible HTML content. They ignored JSON-LD, Microdata, and RDFa entirely. This was a direct-fetch test, which doesn’t prove schema has no role during training, indexing, or retrieval phases.

But it does show that when AI systems grab content from your page to answer a query, they’re not looking at your schema markup.

What this means for your AI visibility strategy

Schema markup gets recommended frequently for AI visibility. Ahrefs’ data complicates this advice. While schema supports rich results and knowledge graphs in traditional search, adding JSON-LD doesn’t increase AI citations for pages that are already cited.

The data shows a correlation between schema and AI citations, but Ahrefs interprets this as a sign of overall site quality rather than schema’s direct impact. Sites that implement schema tend to do other things well too.

This doesn’t mean schema is worthless. It still helps with traditional search features. But if your goal is specifically improving AI citations, schema markup might not be the answer.

Better approaches for improving schema markup AI citations

If schema doesn’t directly boost AI citations, what does? The answer isn’t fully clear yet, but the research points to content quality and authority signals.

Pages that get cited by AI tend to be from sites that invest in comprehensive content, earn quality backlinks, and maintain technical excellence. Schema might be one signal in a broader pattern of quality, but it’s not the determining factor.

Focus on creating content that directly answers the questions people ask AI systems. Make sure your pages load quickly and are technically sound. Build authority through quality backlinks and consistent publishing.

The research gap that still needs filling

The biggest question remains unanswered: does schema markup help pages that aren’t currently cited by AI get their first citations?

This study focused on pages already in the consideration set. For pages struggling to get noticed by AI systems at all, schema might still play a role in initial discovery and indexing.

A different study would need to track pages with no AI citations and measure whether adding schema helps them break through. This is a harder experiment to design and would require a longer measurement period.

Until someone runs that test, we’re left with educated guesses about schema’s role in initial AI discovery.

The current data is clear: if your pages are already being cited by AI, adding schema markup won’t boost those citation rates. Your time is better spent on content quality and technical fundamentals.

For businesses serious about AI visibility, tracking your current citation performance becomes essential before making optimization decisions. AI Mentions helps identify which specific queries trigger competitor citations instead of yours, revealing exact content gaps that prevent AI systems from recommending your brand. You can explore how AI Mentions diagnoses citation gaps beyond simple tracking metrics.