Google retired FAQ rich results in May 2026. Here's what that actually means for FAQ schema and AI search citation, honestly.
FAQ schema, technically FAQPage structured data, is JSON-LD markup that explicitly labels a page's questions and their corresponding answers so search engines and AI systems can identify the question-answer relationship in the content. The honest 2026 answer to whether it helps AI visibility is more nuanced than most guides suggest: Google officially retired FAQ rich results, the expandable dropdown display in search results, on May 7, 2026, and Google's own generative AI search guidance states plainly that structured data isn't required for AI Overviews or AI Mode, with no special schema needed for either.
FAQPage schema itself is not deprecated, and Google has said unused structured data doesn't cause problems for Search, so there's no need to strip it out. What changed is the visible SERP treatment: the expandable Q&A dropdown that used to appear beneath a search listing no longer renders for any site, following an earlier restriction in 2023 that limited it mostly to government and health domains. Search Console's FAQ reporting and the Rich Results Test lose FAQ support in June 2026, with the API losing it in August 2026.
This is genuinely contested, and worth hedging honestly rather than repeating a confident claim either way. Some 2026 analyses report FAQ-schema pages are meaningfully more likely to appear in Google AI Overviews and to be cited by ChatGPT, but that claim isn't confirmed by Google or any AI vendor directly, and should be treated as an unproven correlation rather than an established fact. What Google has confirmed is narrower: structured data isn't a required input for its generative AI features, though schema that accurately matches visible page content is still a reasonable, low-risk practice.
The practical guidance that holds up regardless of the schema debate: only mark up FAQ content where a page genuinely contains real questions people ask, with real, standalone answers, not FAQ blocks added purely to chase a rich result that no longer exists. Structured data must match what's visibly on the page, this is a hard rule for a reason, mismatched schema is a trust problem, not a shortcut. The value of a genuinely useful FAQ section, clear questions, concise 40 to 100 word answers, doesn't depend on whether Google renders a dropdown for it.
This is part of AI Search & SEO. For the broader technical picture, see structured data and llms.txt for AI search visibility. See genuine, question-led content structure applied in AI visibility: from invisible to citable.
Not sure whether your FAQ content is pulling its weight? Get in touch.