A study of 1,885 pages adding structured data found almost no impact on AI overview citations. The bigger finding is what actually drives AI visibility.
Schema markup is good technical hygiene. It's not an AI visibility strategy.
The Take: Ahrefs tracked 1,885 pages that added schema markup and found no meaningful lift in AI citations. Original data and genuine expertise predict citations. Clean code does not.
The SEO industry has been telling businesses to add schema markup to improve their AI visibility. Ahrefs just tested that theory at scale. The results are uncomfortable.
Does Adding Schema Markup Actually Improve AI Citations?
Ahrefs tracked 1,885 pages that added JSON-LD schema between August 2025 and March 2026, matched against 4,000 control pages, and measured citation changes across Google AI Overviews, AI Mode and ChatGPT. Adding schema produced no meaningful lift on any platform.
Ahrefs put a number on the uncertainty: the odds of the Google AI Overviews result being pure chance sit at roughly one in 2,500. That's the one platform where the movement was real, and it moved backwards.
That's a problem for every SEO consultant who has been selling schema implementation as an AI readiness strategy.
Why Do Pages With Schema Still Get Cited More Often?
Pages already cited by AI are almost three times more likely to carry schema than pages that aren't, but that's correlation, not proof schema causes citations. Sites careful enough to add structured data are usually also the ones investing in original content, authoritative writing and technical SEO generally. The schema rides along with the real signal. It isn't the signal.
A separate Ahrefs analysis from the same research programme found that top-10 Google citations in AI Overviews dropped from 76% to 38% over the study period. AI systems are increasingly pulling from sources outside the traditional search rankings, and pages ranked beyond position 100 now supply roughly as many citations as the top 10 does.
The drop in AI overview citations coming from top-10 Google search results
This means the old playbook of ranking well in Google and assuming AI citations would follow is breaking down. The platforms generating AI overviews are building their own quality signals, and those signals don't map neatly onto the signals that drive traditional search rankings.
For Australian businesses investing in SEO, this raises a practical question: where should the effort go?
The answer isn't to stop implementing schema markup. Structured data still helps search engines understand your content, supports rich results and improves how your pages appear in traditional search. Those are real benefits. Keep doing it.
But if your SEO strategy for 2026 is built around the assumption that better structured data equals more AI visibility, you need to rethink the whole plan.
What the research suggests actually drives AI citations is more fundamental. Original data that can't be found elsewhere. Clear, authoritative statements on specific topics. Content that directly answers questions rather than optimising around keywords.
This is why businesses with proprietary datasets, original research or genuine subject matter expertise have a structural advantage in AI search. The machines aren't looking for the best-optimised page. They're looking for the most useful answer.
The businesses winning in AI search aren't the ones with the cleanest markup. They're the ones with the most original thinking.
We built our own content strategy on exactly this bet. Our methodology behind the New Rebellion benchmark dataset, scores across Australian businesses and industries, exists precisely because no one else has it. That's not a schema tag. It's a moat. See how the scoring works.
The tactical takeaway: keep your schema markup clean, but don't treat it as your AI strategy. Invest instead in content that contains information AI systems can't find anywhere else.
Frequently asked questions
Should I stop adding schema markup to my website?
No. Keep it clean and current. It still helps traditional search engines understand your content and supports rich results. It just isn't the lever that moves AI citations, so don't budget for it as an AI visibility strategy.
What actually predicts whether a page gets cited by AI?
Original data, clear authoritative statements and content that answers a specific question directly. Ahrefs found pages with schema get cited more often mainly because the same sites tend to also produce that kind of original, well-maintained content, not because of the markup itself.
Does ranking first in Google still guarantee an AI citation?
Not anymore. Top-10 Google rankings supplied 76% of AI Overview citations in mid-2025. By 2026 that had dropped to 38%, with pages ranked beyond position 100 supplying a similar share to the top 10. Ranking well helps, but it no longer guarantees a citation.