Google killed FAQ rich results on May 7, 2026. Most SEO teams saw the announcement and started removing FAQ schema from their pages. They are making a costly mistake. The rich result display feature is gone but the underlying FAQPage markup is now one of the highest-value AI visibility signals you can implement. Pages with FAQPage schema are 3.2x more likely to appear in Google AI Overviews. ChatGPT, Perplexity, and Gemini all use FAQ-structured content as a primary citation signal. The teams scrapping their schema are handing a visibility advantage to their competitors.
What Google Actually Removed on May 7, 2026
The deprecation announcement is precise about what changed, and most coverage is getting it wrong.
What Google removed: the visual FAQ expansion in search results, the expandable question-and-answer dropdowns that appeared beneath organic listings. That display feature no longer exists for any website. Government and health sites that had retained access since the 2023 restriction lost it too. The FAQ rich result report in Search Console disappears in June 2026. The Search Console API support follows in August.
What Google did not remove: the FAQPage schema type itself. Google's own documentation is explicit. FAQ structured data can stay in place without causing problems, and Google will continue using it to better understand pages. The company stated this directly in the deprecation notice. FAQPage remains a valid Schema.org type, still crawled and processed by Google, Bing, PerplexityBot, and every retrieval-augmented generation crawler indexing the open web.
The distinction matters enormously. A rich result is a display feature. A schema type is a communication layer between your content and machines. Google ended the former. The latter is more useful than it has ever been.
FAQPage schema is structured markup that communicates question-and-answer content to machines in a format they can parse without interpretation. Google ending the visual rich result has no effect on how AI retrieval systems read and use this markup.
The Mistake 80% of SEO Teams Are About to Make
This is not the first time this has happened. In August 2023, Google restricted FAQ rich results to government and health websites only. Most sites lost the feature overnight. The reaction was predictable: SEO teams across the industry started removing FAQ schema from their pages. The logic seemed reasonable. If it does not produce a rich result, why maintain it?
That reasoning was wrong in 2023 and it is even more wrong now. The teams that removed FAQ schema after the 2023 restriction quietly lost a layer of AI citation signal right as AI Overviews began expanding. They were optimizing for a display feature while ignoring the AI infrastructure layer running underneath it.
The May 2026 deprecation will trigger the same reaction at scale. Developers will open their Schema Markup Validator, see FAQ schema flagged as producing no rich result, and remove it in a batch update. Some will do it because a plugin auto-flags it. Others because an agency recommendation says to clean unused schema. The result will be the same: pages that could have appeared in AI-generated answers will stop appearing.
Do not remove FAQ schema. If anything, add it where it is missing.
Why FAQ Schema Is AI Fuel, Not a Display Feature
The reason FAQ schema matters for AI search has nothing to do with rich results and never did. Large language models construct answers in question-and-answer format. When someone asks ChatGPT what is the best tool for tracking brand mentions in AI search, the model is looking for content it can extract as a direct answer to a specific question. FAQ-structured content is pre-formatted exactly the way AI systems prefer to retrieve and present information. The question labels the intent. The answer provides the extractable response. The model does not have to guess where one ends and the other begins.
Google's own AI Overviews system draws heavily on structured data to identify, extract, and cite precise passages from web pages. A properly implemented FAQPage schema provides semantic anchors that make your content easier for the AI Overview generation pipeline to extract and cite. The AI Overview does not need to infer what your content is answering. The schema tells it directly.
Beyond Google, Microsoft confirmed at SMX Munich 2025 that schema markup helps its LLMs understand content for Copilot. PerplexityBot crawls and processes FAQPage markup. Every major AI retrieval system is reading these signals. The brands that maintain comprehensive FAQ schema are building a semantic layer that compounds across every AI platform simultaneously.
One important nuance: LLMs like ChatGPT tokenize JSON-LD as raw text rather than parsing its semantic structure. The direct benefit of FAQ schema for AI citation comes from two pathways. First, the visible on-page Q&A content that LLMs extract directly. Second, the Google Knowledge Graph pathway where schema strengthens entity understanding that AI systems reference. Both matter. Neither requires a rich result to function.
The 3.2x Multiplier: What the Data Shows
The most important number in this conversation is 3.2. Pages with FAQPage markup are 3.2x more likely to appear in Google AI Overviews compared to pages without FAQ structured data, according to analysis published by Frase.io and corroborated by independent research across multiple datasets. A separate Relixir study analyzing 50 sites found pages with FAQPage schema achieved a citation rate of 41% versus 15% for pages without it.
Google AI Overviews now appear on approximately 31% of all search result pages. Position one organic CTR has dropped 32% year over year as a direct result. The channel that used to reward your SEO investment is shrinking. The channel that FAQ schema directly feeds is growing.
If your page already ranks in Google's top 10 for a keyword, adding FAQ schema increases your probability of appearing in AI Overviews for that query by approximately 40%. That creates dual visibility: a traditional organic listing and an AI-generated citation in the same SERP. Removing FAQ schema eliminates one of those two appearances entirely.
The brands that understood this after the 2023 restriction maintained their FAQ schema and have been compounding that advantage for three years. The brands that removed it have been invisible in AI Overviews during one of the fastest growth periods AI search has ever seen.
How ChatGPT, Perplexity and Gemini Use FAQPage Markup Differently
Not all AI systems interact with FAQ schema through the same pathway, and understanding the differences changes how you implement it.
Google AI Overviews
Google AI Overviews use FAQPage schema as a direct semantic signal. When the crawler processes your page, FAQ markup provides explicit anchors that make your content more extractable for the AI Overview generation pipeline. This is the pathway behind the 3.2x multiplier. Google's systems read the schema, understand the question-answer structure, and factor it into citation decisions.
ChatGPT
ChatGPT does not parse JSON-LD semantically. A February 2026 controlled experiment confirmed that ChatGPT tokenizes structured data as raw text, which means the schema structure itself is not read the way Google's crawler reads it. For ChatGPT, the value of FAQ schema is indirect: it strengthens your representation in Google's Knowledge Graph, which influences which pages appear in the web search results ChatGPT references. The direct citation driver for ChatGPT is the visible on-page Q&A content, short and direct answers structured as explicit question-and-answer pairs in your HTML.
Perplexity
Perplexity operates a hybrid model. PerplexityBot crawls FAQ markup directly and uses it as a retrieval signal. Perplexity also pulls heavily from recent content and conversational sources. FAQ schema on a well-structured page gives Perplexity's retrieval system cleaner extraction pathways, particularly for answer queries where the response needs to be synthesized from a single authoritative source.
Gemini
Gemini follows patterns closest to Google AI Overviews, given shared infrastructure. FAQ schema feeds into the same semantic understanding layer that powers AI Overview citation decisions. Implementing FAQ schema correctly for Google essentially means implementing it correctly for Gemini at the same time.
How to Implement FAQPage Schema Correctly in 2026
Step 1: Write the FAQ content visibly on the page first
The schema must match content that users can actually see. Do not add FAQPage markup for questions that only exist in the JSON-LD. Every question and answer in your schema must be visible in the HTML. This is a Google requirement and it is also what LLMs extract directly. Each answer should be between 40 and 60 words. This length maps directly to how LLM retrieval systems chunk content. Self-contained answers in this range fit within a single retrieval chunk, which means the model can extract a complete, coherent response.
Step 2: Use JSON-LD format in the page head
JSON-LD is the recommended format. It sits cleanly in a script tag separate from your HTML, which makes it easier for crawlers to parse and does not interfere with your visible content structure. Use the @type FAQPage with mainEntity as an array of Question objects, each containing an AcceptedAnswer. Google's official documentation at developers.google.com/search/docs/appearance/structured-data/faqpage has copy-ready examples.
Step 3: Match the schema question text exactly to your visible H3 heading
Structure your FAQ questions as H3 headings in your visible content. The text in your schema's name property should match the heading text exactly. This consistency helps AI platforms verify the relationship between the markup and the content and increases extraction confidence.
Step 4: Stack FAQPage with Article schema
Nesting FAQPage inside an Article schema creates a compound signal. You are telling AI systems both the content type and the specific Q&A pairs it contains. Research shows pages with three to four complementary schema types get cited at twice the rate of pages with a single schema type. At minimum, implement Article and FAQPage together on content pages.
Step 5: Keep answers specific and data-driven
Weak answer: FAQ schema is important for AI search visibility. Strong answer: FAQ schema increases citation probability by 3.2x in Google AI Overviews and achieves a 41% citation rate across AI platforms versus 15% for unstructured content, according to Frase.io and Relixir research published in 2025 and 2026. AI platforms cite specific claims with data behind them at significantly higher rates than vague assertions. Every FAQ answer should contain at least one concrete, verifiable fact.
Step 6: Validate and monitor
Use Google's Rich Results Test to confirm your JSON-LD is technically valid. Note that from June 2026 onward, Search Console will no longer report FAQ rich result impressions. To know whether your FAQ schema is actually driving AI citations, you need a monitoring layer that tracks your brand mentions across AI engines directly.
Frequently Asked Questions
Does removing FAQ schema after Google's May 2026 deprecation make sense?
No. Google deprecated the FAQ rich result display feature, not the FAQPage schema type. The markup continues to be processed by Google for page understanding, by Bing and Copilot, and by AI retrieval systems including PerplexityBot. Removing it eliminates an AI citation signal without any ranking benefit.
How does FAQ schema help with Google AI Overviews specifically?
Pages with valid FAQPage markup are 3.2x more likely to appear in Google AI Overviews compared to pages without it. The schema provides semantic anchors that make your content easier for Google's AI to extract and cite. AI Overviews now appear on approximately 31% of all search result pages, making this multiplier increasingly significant.
Does FAQ schema directly improve ChatGPT citations?
Not directly. ChatGPT tokenizes JSON-LD as raw text without parsing its semantic structure. For ChatGPT, the value is indirect: FAQ schema strengthens your Google Knowledge Graph representation, which influences the web results ChatGPT references. The direct driver for ChatGPT citations is well-structured visible Q&A content on the page itself.
What is the correct answer length for FAQ schema in 2026?
Between 40 and 60 words per answer. This length maps to how LLM retrieval systems chunk content. Answers in this range fit within a single retrieval chunk, meaning the model extracts a complete, coherent response. Shorter answers lack substance. Longer answers risk being split across chunks and losing coherence in the extracted output.
Should I add FAQ schema to every page on my site?
Only add FAQPage schema to pages that genuinely contain question-and-answer content visible to users. Thin FAQ sections added solely for schema purposes violate Google's guidelines and provide low citation value. Prioritize your most important product, service, and pillar content pages where real questions about your offering are already answered.
See Exactly Where Your FAQ Schema Is Driving AI Citations
Most teams implement FAQ schema and then have no way to know whether it is actually working. Google Search Console reported rich result impressions, but from June 2026 onward, FAQ data disappears from those reports entirely. You are left guessing.
Vizible AI tracks your brand mentions daily across ChatGPT, Claude, Gemini, Perplexity, Mistral and DeepSeek. You can see exactly which pages are getting cited, which queries trigger mentions of your competitors instead of you, and whether your structured data changes are translating into real AI visibility improvements. The citation gap view shows you the prompts where competitors appear and you do not, which becomes your content and schema priority list.
Want to know if your FAQ schema is helping you get cited in AI answers? Vizible AI shows you exactly where your brand appears and where it does not. Try it free at vizibleai.com