Vendor reportAhrefs, 11 May 2026

Adding JSON-LD schema to already AI-cited pages did not lift citations on Google AI Mode or ChatGPT, and their AI Overviews citations fell slightly against matched pages

Optimisation methods

AI source	Effect on citations (%)
Google AIO	-4.6%
Google AI Mode	2.4%
ChatGPT	2.2%

Effect of adding JSON-LD schema on AI citation rates by platform

Reader Note:: The test found no lift. And because every page studied already carried AI Overview citations before schema was added, it says nothing about pages AI has not found yet, which is where most sites sit. A smaller visibility test published by Search Engine Land pointed the other way, arguing that schema quality helps a page appear in an AI Overview, but it never says how many pages it ran on or how large the effect was.
Evidence:: Experimental
Vendor interest:: Moderate — This comes from a company with a commercial interest in this market, so it is not a neutral source.
Sample:: Main study: 1,885 web pages that added JSON-LD schema between August 2025 and March 2026, matched to 4,000 control pages from different domains with similar pre-period citation levels that never added JSON-LD. All sampled pages already had 100+ AI Overview citations in February 2025.

Good for

Backing budget and pricing decisions in similar situations.

Not for

Situations far from the one tested, without checking again.
This comes from a company with a commercial interest in this market, so it is not a neutral source.
This is an aggregated or average number that covers very different cases. Real results in a particular segment may differ.

Structure boosts citation odds: clean heading hierarchies, FAQs, and schema earn citations at 2 to 2.8 times the rate of unstructured pages

Structural attributes: ChatGPT-cited pages vs Google-only top SERPs
Attribute	ChatGPT-Cited (%)	Google Only (%)
Single H1 Present	86.8%	64.2%
Lists Sections Exist	78.3%	26.8%
Proper Heading Alignment	68.7%	23.9%
Rich Schema	60.5%	24.5%

Changing badge strength (urgency, discount, or scarcity) usually did little for AI recommendations, while the star score was the clearest exception — and the only change that actually boosted recommendations

Go deeper on this topic

Our Substack turns the data behind these cards into full breakdowns: what changed, why, and what to do about it.

Fixing your influencer marketing campaign: How to get into the top 10% of advertisers

Product ratings are the only badge that helped every AI model in every product category to get recommended more often

Self-promotional listicles can help short-term in AI search but are already being targeted for demotion by Google.

Optimizing for Google organic search remains the main way to get cited in major AI search products.

16.39%

of ChatGPT retrieved pages are product pages, now more than listicles

Adding JSON-LD schema to already AI-cited pages did not lift citations on Google AI Mode or ChatGPT, and their AI Overviews citations fell slightly against matched pages

Good for

Not for

Related cards

Go deeper on this topic