Google Killed FAQ Rich Results in May 2026. Here’s the Truth About Your FAQ Schema.

On May 7, 2026, Google stopped showing FAQ rich results in Google Search. The visible SERP feature is gone. FAQPage structured data is still a valid Schema.org type, and Google states the markup can stay on your pages without causing problems. Remove it only where the Q&A was stuffed, duplicated, or fake. Keep it where the questions are real. Do not believe anyone — including Google’s own documentation — who tells you the markup is now a confirmed AI-ranking lever. It isn’t. The reliable value is the clear, visible Q&A content itself.

That is the whole decision. Everything below is the evidence, the parts the industry is getting wrong, and a keep/fix/remove matrix you can run today.

Key facts at a glance

What Google actually changed, on three dates

Google removed a display feature. It did not remove a data format, and rich results were never a ranking signal. The change rolls out in three stages, all stated in a deprecation note at the top of Google’s FAQPage developer documentation.

May 7, 2026 — the visible block is gone. Standard Search results no longer show the expandable FAQ accordion beneath organic listings. Pages that gained extra vertical space from that block now display like any other result. Click-through rate can fall on pages that relied on that larger footprint. Rankings do not change, because the rich result was a presentation layer, not a ranking input.

June 2026 — reporting and testing drop FAQ. The FAQ search appearance filter, the FAQ rich result report in Search Console, and FAQ validation inside the Rich Results Test are removed. Historical data remains for the standard retention window; new FAQ-specific impression and click reporting ends.

August 2026 — the Search Console API stops returning FAQ data. This is the date most coverage omits. If you pull FAQ rich result metrics into dashboards or BigQuery through the Search Console API, those calls return empty after August 2026. Export anything you need first and update the queries.

What did not change: FAQPage remains a syntactically valid Schema.org type. Google’s documentation states the markup can stay without causing problems. The feature is gone; the format is not.

What Google did say — and what it did not

This section exists because the most-shared takes misquote Google.

Google did say: the rich result is gone on the dates above, and that you can leave the markup in place. That is the full extent of Google’s statement.

Google did not say the markup helps AI Overviews, ChatGPT, Gemini, or any AI system. There is no line in Google’s FAQ documentation claiming a generative-AI or “page understanding” benefit from FAQPage. Coverage asserting “Google confirmed it will keep using FAQ data to understand pages” is adding words Google did not write. Treat that as industry inference, not a Google statement.

Context that is true: FAQ rich results had already been restricted since 2023 to well-known authoritative government and health sites, after widespread abuse — sites adding artificial FAQ blocks purely to occupy more SERP space. The May 2026 change ends the feature for those remaining sites too. For the long tail, the rich result had been functionally dead for roughly three years.

Should you remove your FAQPage schema?

In most cases, no. Removing valid FAQPage markup neither helps nor harms Google rankings, and Google says unused structured data does not create Search problems. The only markup worth removing is the markup that was already a problem: keyword-stuffed questions, near-duplicate FAQ blocks repeated across pages, or questions no real user would ask. Note that Google’s own guidelines already say repetitive FAQ content should be marked up on only one page sitewide — many sites violated this for years to inflate footprint.

Run this matrix across the whole site before touching anything.

Does FAQ schema still help in ChatGPT, Gemini, and AI Overviews?

Honest answer: the clear Q&A content helps; the markup’s direct effect is unproven and probably small. Two opposite overcorrections are circulating — “schema is dead” and “FAQ schema matters more than ever for AI” — and neither is accurate. Here is the defensible middle.

How AI systems actually read pages. ChatGPT, Gemini, Perplexity, and Copilot retrieve and chunk the rendered text of a page. They generally tokenize JSON-LD as raw text rather than parsing it as a structured graph the way Googlebot does. So a question followed by a short, declarative answer is easy to extract and cite — but that is a property of the visible content, not of the markup wrapper.

Where structured data does help indirectly. It helps Google’s own pipeline classify entities and content, which can feed Google’s surfaces. It also keeps your visible content and your machine-readable description consistent, which reduces ambiguity. Those are real but indirect benefits.

What is contested. Whether the FAQPage markup itself causally increases AI citations is disputed. Vendor studies report large multipliers — figures from roughly 25% to 40% higher citation probability, and 2x to 3x multipliers, appear across different tools — but they use different query sets, platforms, and methods, and most do not control for the fact that schema-bearing pages also tend to have clearer visible content. Treat any specific “Nx more citations” number, including ones we could generate from our own client data, as directional and uncontrolled, not proof of causation.

Practical conclusion. Keep substantive FAQ content because clear, self-contained Q&A is extractable and citable by every engine. Keep the schema because it is valid, free, and may help indirectly. Do not promise a stakeholder a citation multiplier you cannot defend with a stated methodology.

A clean FAQPage implementation (the format worth keeping)

Use JSON-LD. One concept per question. Lead each answer with a single declarative sentence, then add brief support. Ensure every question and answer is also visible on the page — Google requires the content be user-visible, and AI engines extract the visible text.

json

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "Did Google remove FAQ rich results?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Yes. As of May 7, 2026, FAQ rich results no longer appear in Google Search. The FAQPage markup remains a valid Schema.org type and can stay on the page."
      }
    },
    {
      "@type": "Question",
      "name": "Should I delete my FAQPage schema?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Only if the Q&A is stuffed, duplicated across pages, or covers questions no real user asks. Valid, substantive FAQ markup is harmless to keep and does not affect rankings."
      }
    }
  ]
}

The visible content on the page must mirror this markup exactly.

What to build instead of chasing the next SERP feature

The reflex after a deprecation is to hunt for the next feature. The durable move is content that survives the next five changes.

Structure each Q&A for extraction. One concept per question. A single declarative opening sentence that fully answers it. One to four short supporting sentences. No hedging, no keyword stuffing.

Match visible content to schema. Search and AI systems compare rendered HTML against structured data; mismatches reduce trust. Whatever is in the markup must appear on the page.

Add Person and Organization signals. A verified author (Person) and clear publisher (Organization) raise confidence, especially for health, finance, legal, and technical topics that Google holds to higher standards.

Cross-link Q&A to pillar pages. Each answer should link to a relevant pillar or service page to reinforce topical authority.

Cite credible sources in your content. AI systems favor well-sourced, corroborated content. Original data helps only when you publish the methodology.

Replace FAQ rich-result reporting with citation tracking. Search Console FAQ metrics end in stages through August 2026. Tools such as Otterly.ai, Profound, and Peec.ai monitor whether your content is cited inside AI responses across ChatGPT, Gemini, Perplexity, and Copilot. Move measurement there before the data goes dark.

Why this was predictable

Google did not act on impulse. The 2023 restriction limited FAQ rich results to authoritative government and health sites in response to abuse. From that point the feature was effectively gone for the long tail. The May 2026 change finishes the process and ends eligibility for the remaining sites.

The direction is consistent: Google is reducing presentation formats that compete with AI surfaces for space. Any single SERP feature you build a strategy around is a candidate for the same treatment. HowTo rich results followed the same pattern FAQ did — restricted in 2023, then wound down — which makes HowTo the most plausible next deprecation. That is informed speculation, not a Google announcement.

How the impact differs by market

Treat the figures below as ranges. AI Overview prevalence varies widely by study, geography, device, and method.

United States. AI Overviews already appear on a large, growing share of US queries — Google has described it as roughly half of US searches, while independent trackers report anywhere from about 20% to 50% depending on method, and informational queries trigger them most often. Seer Interactive’s analysis put informational-query coverage near 36%. The FAQ block was already losing clicks to the AI Overview above it, so the deprecation accelerates a shift already underway. Expect little direct traffic loss from the deprecation itself; the larger ongoing pressure is AI Overview displacement.

Australia. AI Overview coverage in AU has trailed the US, and FAQ rich results reportedly persisted in AU SERPs somewhat later. Sites that built around the visible block may feel the change more sharply, but they also have a wider window to rebuild around AI-citation strategy before the gap closes.

LATAM and Spanish-language markets. AI search coverage in Spanish has lagged English, and rich results were less consistently distributed. The deprecation is closer to a non-event short term. The medium-term opportunity is to ship clean, AI-aligned FAQ content before competitors do, building citation positions that compound through 2026.

Schema priority stack for May 2026

FAQPage stays high — not because the markup is a proven AI lever, but because the underlying clear Q&A content is, and the schema costs nothing to retain.

 Frequently Asked Questions About Slow Business

Author: Adriana Kligman, Founder and CMDO of Hyperdot. Google Partner. Meta Partner. Architect of the firm’s GEO methodology since 2016Author: Adriana Kligman, Founder and Chief Marketing Officer of Hyperdot. Google Partner. Meta Partner. Leading the firm’s search and structured-data practice since 2016, and its generative engine optimization (GEO) practice since the discipline emerged. LinkedIn