Google Retired FAQ Rich Results. Keep the Q&A, Drop the Markup Obsession.

If you still have FAQPage schema on your site and you're wondering whether to rip it out, breathe. The decision is simpler than the noise suggests — but it does require understanding what Google actually did, and what didn't change alongside it.

The short version: faq rich results are gone from Google Search as of May 7, 2026. The expandable accordion panels that once appeared below your blue link — listing out questions with answers readers could click open — will not come back. But the underlying markup, the content strategy, and the AI citation opportunity around Q&A? Those are very much alive, just pointing in a different direction.

What Google Actually Removed (and When)

This wasn't a single announcement. It was a three-stage retirement that played out over nearly three years:

• August 2023: Google restricted FAQ rich results to "well-known, authoritative government and health websites" only. Commercial sites, agencies, and blogs lost eligibility immediately. Most sites went three years without seeing a single FAQ snippet after this date.
• May 7, 2026: FAQ rich results stopped showing in Search entirely — including for the government and health publishers who had kept them since 2023.
• June 2026: The FAQ rich result report in Search Console was retired, and support was removed from the Rich Results Test tool.
• August 2026: Search Console API support for FAQ rich result data was removed — the final stage of the phaseout.

The official record is a documentation update added on May 8, 2026 — not a dedicated blog post or major announcement. Google quietly added a deprecation banner to its FAQPage developer documentation. Within hours, half the SEO industry declared FAQ schema dead. The reality was considerably less dramatic.

Why Google Retired FAQ Rich Results

The honest answer combines two factors: abuse and strategic direction.

On the abuse side, FAQ markup had gotten out of hand. Marketers bolted generic FAQ blocks onto nearly every page to grab extra real estate in search results, whether the content deserved it or not. Google engineer Ryan Levering described the situation as FAQ markup looking "like a spam vector" — in the same category as fake review stars. The 2023 restriction was the cleanup. The 2026 retirement simply closed the loop.

On the strategic side, the pattern is hard to miss. HowTo rich results were retired from desktop in September 2023. Seven more structured data types — including Book Actions, Course Info, and Claim Review — were retired in June 2025. In every case, the surface that gained ground was Google's own AI Overview. Google is reclaiming SERP real estate for its own answer experiences, not penalizing structured data itself.

The fairest reading: Google is pruning low-value visual treatments, not declaring structured data irrelevant.

The Crucial Distinction: Markup vs. Rich Result

This is where most of the confusion lives, so it's worth being precise.

FAQPage schema is a type of structured data. It tells search engines that a section of content on your page is organized as questions and answers. FAQ rich results were a SERP display feature that used that markup to render an expandable panel beneath your search listing.

Those are two separate things. Google ended the display feature. The markup type remains valid and supported. Google has explicitly confirmed that FAQPage markup will not trigger a manual action, will not damage rankings, and still helps machines parse your content structure. There is no reason to mass-delete it from your site.

What you should stop doing: treating FAQ schema as a visibility lever. It no longer moves any needle in the traditional SERP. Add it where Q&A content genuinely exists on the page; don't add it hoping for a boost that isn't coming.

Where FAQ Markup Still Does Real Work

The presentation layer ended. The comprehension layer didn't.

AI citation: the new primary value

Here's what hasn't changed at all: AI engines — Google AI Overviews, ChatGPT, Perplexity — actively look for structured, parseable content when deciding what to cite in generated responses. Q&A content is one of the cleanest formats they can extract.

The structured Q&A format is particularly powerful because it mirrors the exact format AI systems use internally — question and answer pairs. When you mark up a section with FAQPage schema, you're giving AI crawlers an unambiguous map: "this is the question; this is the authoritative answer."

Google's May 2026 generative-AI guidance states that "structured data isn't required for generative AI search," which is technically true. But the visible Q&A content the schema mirrors is still directly extractable by every major AI platform. The schema labels it clearly; the content does the actual work.

Bing and other surfaces

Google retiring the FAQ rich result does not bind Bing or other engines. Bing Webmaster Tools still validates FAQPage markup, and Microsoft Copilot uses structured Q&A signals when generating responses. For any store or publication with meaningful Bing traffic, keeping valid FAQ markup is low-cost insurance. As we've covered in our piece on zero-click search, the search landscape is distributing across more surfaces, not fewer — and that makes schema hygiene matter more, not less.

Content comprehension signals

Even setting AI citation aside, labeled Q&A content is easier for machines to extract accurately than unstructured prose. Schema acts as a translation layer between your content and the machine: it removes ambiguity, defines relationships between entities, and confirms that what's visible on the page matches what's in the code. That consistency check matters — AI systems actively look for it.

What Actually Earns Citations Now: Writing FAQ Content That Gets Quoted

The FAQ rich result rewarded having the markup. AI citation rewards the quality of the writing. That's a meaningful shift in where effort should go.

Here's what the research and real-world testing point to for FAQ answers that get cited in AI-generated responses:

• 40–60 words per answer. That range is long enough to fully resolve a question and short enough for an AI engine to quote as a complete unit. Answers under 40 words often lack sufficient context to be cited; answers over 60 words dilute precision and reduce the chance an AI engine selects your exact passage.
• Lead with the direct answer. Don't build to the answer — open with it. AI systems using retrieval-augmented generation (RAG) extract passage chunks without surrounding context. Your answer must make complete sense in isolation, without the question above it or the article around it.
• Keep questions under 15 words. Longer questions fragment poorly when AI systems decompose them into sub-queries. Match the phrasing your actual customers use — pull from support tickets, search query reports, and "People Also Ask" results for your category.
• 3–6 FAQs per page is the optimal range. Fewer than three gives AI crawlers insufficient extraction targets. More than six dilutes signal quality — AI engines don't process every entry, they select the most relevant Q&A pair for a given query.
• Every FAQ answer must appear as visible text on the page. Markup that doesn't match visible content is bad practice and was always against Google's guidelines. That principle hasn't changed.

This is also exactly why AI citation tracking has become a more important metric than FAQ rich result impressions ever were. If you're not measuring where your content gets cited across AI platforms, you're flying blind. Our guide on tracking AI visibility walks through the tools and methods for doing that in 2026.

The 10-Minute FAQ Schema Audit

If you want to clean up your setup without overthinking it, here's what to do right now:

1. Check Search Console for lingering FAQ reports. If you're still seeing FAQ rich result data in your reports, that's stale. The report was retired in June 2026. If your dashboard still shows it, your API integrations or third-party reporting tools may need updating to avoid silent failures and stale data.
2. Scan your top 20 pages for FAQPage schema. Use the Rich Results Test (note: it no longer validates FAQ, but Schema.org's validator at validator.schema.org still checks for errors). Flag any pages where the schema exists but there's no visible Q&A content on the page — that's the one scenario worth cleaning up.
3. Audit the question quality on each page. Replace generic filler questions ("What is your return policy?") with questions your actual customers ask. Pull from your search query report, your support inbox, and the "People Also Ask" box for your main keywords.
4. Rewrite answers to the 40–60 word standard. Each answer should stand alone, open with a direct response, and include one supporting fact or specific detail. Generic answers don't get cited.
5. Redirect structured data energy toward schema types that still earn rich results. Product, Article, Review, Event, and Organization all remain eligible in 2026. If you're a Shopify seller and your Product schema isn't complete and validated, that's a higher-priority fix than anything FAQ-related.

The whole exercise takes about ten minutes per page for the audit, longer if you're rewriting answers — but the rewrite is where the AI citation value actually lives. And it compounds: better-written answers help get cited by AI search engines the same way that better-written articles earn organic links. Quality is the durable asset.

The Bigger Pattern Worth Noticing

FAQ rich results join a growing list of SERP features Google has retired: HowTo on desktop (2023), seven more types in 2025, and now FAQ in 2026. In each case, Google described the change as removing features that weren't adding enough value for users — often because they'd been gamed at scale by SEO tooling.

The direction is consistent. Google is reclaiming SERP real estate for its own AI-generated answer experiences and reducing the display space available to third-party rich results. That's not hostile to publishers — it's just where the product is going.

Building an SEO strategy around any single display feature has always been a fragile bet. The brands that handle this update best won't be the ones who delete FAQ schema fastest. They'll be the ones who use the moment to audit how much of their content strategy was leaning on display features, and rebuild around pages that genuinely deserve to rank, get cited, and earn clicks on their own merits. Substance survives. Tactics don't.

The same Q&A discipline that feeds AI citation, the same careful writing that makes an answer worth quoting, is also what makes content hold up across different AI platforms that don't share citation sources. That's the real lesson buried inside the FAQ rich results retirement.

Frequently Asked Questions

Do I need to remove FAQ schema from my pages now that rich results are gone?

No. Google explicitly confirmed there is no penalty for leaving FAQPage markup in place. The schema is still valid, still helps machines understand your content structure, and costs nothing to keep. Simply stop treating it as a visibility tactic — it no longer triggers any SERP feature.

Will keeping FAQ schema help me appear in Google AI Overviews?

Indirectly, yes. Google's May 2026 guidance states structured data is not required for generative AI search, but the visible Q&A content the schema mirrors is what AI engines extract and cite. Self-contained answers of 40–60 words with a direct first sentence are the real citation driver.

Which schema types still produce rich results in Google in 2026?

Product, Article, Review, Event, Organization, LocalBusiness, and several others still generate rich results in Google Search. FAQ and HowTo are the notable retired types. Focusing structured data effort on these active types delivers measurable SERP impact.

When did Google first restrict FAQ rich results, and for whom?

Google restricted FAQ rich results in August 2023, limiting them to well-known, authoritative government and health websites only. Commercial sites, blogs, and agencies lost eligibility at that point. The May 2026 update retired the feature entirely, including for the government and health sites that had retained it.

How long should FAQ answers be for AI search citation?

Aim for 40–60 words per answer. That range is long enough to fully resolve a question and short enough for an AI engine to quote as a complete unit. Answers under 40 words often lack context; answers over 60 words dilute precision and reduce the chance an AI selects the exact passage.

Does FAQ schema help with Bing or other search engines besides Google?

Yes. Google retiring the FAQ rich result does not affect Bing or other search engines. Bing Webmaster Tools still validates FAQPage markup, and structured Q&A content remains a signal Bing's Copilot uses when generating responses. Keeping valid FAQ markup is low-cost insurance for non-Google surfaces.