Schema markup alone won't get you cited in AI answers. Google says its AI features need no special markup, Microsoft says schema helps its AI understand pages, and the best-designed study so far found adding JSON-LD made no clear difference to citations. Use structured data because it describes your business accurately and makes pages eligible for rich results. Earn citations the way you earn rankings: pages that answer the question, facts that match everywhere, and mentions on sites the assistants trust.
What Google says
Google's guide to its generative AI features, added in May 2026, is direct about it. Structured data "isn't required for generative AI search, and there's no special schema.org markup you need to add." Google says it's still worth using as part of SEO because it helps pages qualify for rich results, and that AI features in Search are "rooted in our core Search ranking and quality systems." The same guide says Google Search ignores llms.txt and other AI text files.
Google's general structured data guidelines add two rules that matter: markup must describe content people can see on the page, and valid markup never guarantees a rich result.
Some markup no longer does anything visible on Google. HowTo rich results were deprecated in September 2023. FAQ rich results were limited to government and health sites in 2023 and removed from Search in May 2026. Google has said unused markup can stay on the page without harm.
What Microsoft says
Microsoft is more positive. In an October 2025 post on getting content into AI search answers, Bing's Krishna Madhavan lists schema as a core practice because it helps search engines and AI systems understand content, while noting that nothing guarantees selection. Bing Webmaster Tools now reports how often Copilot and Bing's AI summaries cite your pages, so you can watch the effect yourself.
What the studies show
- A matched test found no clear lift. Ahrefs tracked 1,885 pages that added JSON-LD between August 2025 and March 2026 and compared them with matched pages that didn't. Citations in Google's AI Mode rose 2.4% and in ChatGPT 2.2% relative to the controls, neither distinguishable from zero. In AI Overviews the pages did 4.6% worse, and the authors say they can't tell whether schema caused that. The study only covered pages already heavily cited, over 30 days.
- Correlation isn't cause. Semrush's study of 5 million cited URLs found Organization, Article and Breadcrumb were the most common schema types on pages cited by ChatGPT Search and Google AI Mode, and says plainly that this is correlation only. Sites that bother with structured data tend to do other things well too.
All three studies come from SEO software vendors. We found no peer-reviewed test of schema against AI citations.
What to mark up anyway
Use the types that describe what is really on the page, in JSON-LD, which Google recommends:
- Organization on your home or about page, so search engines can tell your business apart from similar names.
- LocalBusiness if customers visit or you serve an area, with hours and address that match your Business Profile.
- Article on posts, with the author and date people can see.
- Product or SoftwareApplication on product pages, and Breadcrumb across the site.
Keep it accurate and current, and spend the rest of the effort on the content itself. Our GEO explainer covers what the evidence says works.
Key Takeaways
- Google says no special markup is needed for its AI features; keep structured data for rich results.
- Microsoft recommends schema for AI answers but promises nothing.
- The best matched study found no clear citation lift from adding JSON-LD.
- FAQ and HowTo rich results are gone from Google; existing markup does no harm.
- Mark up what is really on the page, and put the rest of the effort into content and consistency.