Schema markup is much more common on AI-cited pages. But a new Ahrefs report found that its addition did not result in a clear increase in citations.
Ahrefs tracked 1,885 web pages adding JSON-LD schema. Each page was compared to control pages that never added schema, and citation changes were measured in Google AI Previews, AI Mode, and ChatGPT.
No platform showed a significant increase in citations after adding the schema.
What Ahrefs found
The report analyzed 6 million URLs and found that pages cited by AI were about three times more likely to include JSON-LD. This discrepancy was taken as evidence that the scheme improves AI visibility. However, Ahrefs tested whether this was true when isolated from other signals, as sites with schema tend to invest in better content and earn more links.
They performed a controlled comparison, matching each schema page with three control pages from different domains with similar citation levels that never added JSON-LD. Citation changes were measured 30 days before and after the schema was added.
Using its Brand Radar tool and Agent A, Ahrefs performed a difference analysis to account for platform trends. Here’s what was found.
- Google AI Insights: −4.6% (a slight but statistically notable decrease compared to controls)
- Google AI Mode: +2.4% (too small to be distinguished from random variation)
- ChatGPT: +2.2% (too small to be distinguished from random variation)
Three other tests were run alongside the main comparison, and all four found no clear positive or negative effects.
The decline of AI
THE −4.6% The decline in the AI Overview section deserves context. Ahrefs reporting on processed and control pages was already declining before the schema was added. Pages processed declined a little faster, but the difference is small, with about 12 fewer daily citations per page in a sample where most pages received hundreds.
The report notes that this drop could reflect a slight negative effect of the scheme, or could be a coincidence. This doesn’t draw any conclusions either way.
What the report does not cover
Each page of the dataset contained over 100 AI Overview citations before a schema was added. These pages were already in the consideration set, being explored and surfaced.
The report admits this limitation. For pages that are not yet visible to the AI, schema may still make it easier to crawl, analyze, or index, but the data cannot confirm this.
The report also notes other limitations. Pages adding JSON-LD often modify other elements, making it difficult to separate schema effects from those modifications. All schema types have been grouped together, so some may work differently. The 30-day window might miss slower effects.
A researchVIU experience cited in the report tested whether five AI systems used schema markup when retrieving pages in real time. None did; they only extracted visible HTML, ignoring JSON-LD, Microdata and RDFa. This was a test of direct retrieval, not evidence of the role of schema during training, indexing, or retrieval.
Why it matters
Schema markup is frequently recommended for AI visibility. However, Ahrefs’ data complicates things. Although the schema supports rich results and knowledge graphs, adding JSON-LD does not increase AI citations for already cited pages.
The data shows a correlation: pages with schema are cited more often by the AI, but Ahrefs interprets this as a sign of the overall quality of the site rather than the direct impact of schema.
Looking to the future
The report cannot determine whether the schema helps pages that are not yet cited, which is a different group of pages that requires further study. If the pages are visible to the AI, JSON-LD probably won’t increase citations.
Featured Image: Romain Samborskyi/Shutterstock
