As of May 7, 2026, FAQ rich results no longer appear in Google Search at all, for any site, query type, or language, following a deprecation notice added to Google's FAQ rich result documentation. This is the second and final step in a retreat that began in 2023, when Google first restricted the visible FAQ snippet — the expandable question-and-answer accordion that used to appear directly in search results for a huge range of sites — down to a narrow set of government and health properties. That earlier restriction already eliminated most of the SERP real estate FAQPage schema had originally been implemented to capture. The May 2026 update removes it entirely, for every site, everywhere.
This is a legitimate moment to ask whether continuing to implement FAQPage schema is still worthwhile, and it deserves a direct answer rather than a reflexive "yes, schema always matters" response. The honest answer requires separating two distinct things that got bundled together under one schema type: the visible rich-result feature in classic Google search, and the machine-readable Q&A structure that AI retrieval systems read during citation selection. One of those is now gone. The other is not affected by this deprecation at all.
What Exactly Was Deprecated, and What Was Not?
Google's deprecation notice applies specifically to the FAQ rich result — the visible accordion UI element that used to expand directly in a classic search listing, showing question-and-answer pairs without the user needing to click through to the page at all. That specific SERP feature is gone. Google's own May 15, 2026 generative AI optimization guide, covered in a companion article in this series, separately and explicitly states that FAQPage schema is not required for inclusion in AI Overviews or AI Mode — a statement about Google's generative AI features specifically, distinct from the classic rich-result deprecation, but arriving in the same news cycle and compounding the confusion.
What was not deprecated: the FAQPage schema type itself remains valid, fully supported schema.org markup, and Google's own crawlers continue to parse it without penalty. More importantly for the rest of the AI search landscape, no other major AI engine has issued any comparable statement. ChatGPT, Perplexity, Claude, and Gemini all continue reading FAQPage JSON-LD as a structural signal during retrieval and citation selection, and the independent correlation research cited throughout this series — showing FAQPage schema associated with meaningfully higher citation rates on these engines — was never based on the now-defunct Google rich-result feature in the first place. That research measured citation behaviour in generated AI answers, a different mechanism entirely from the classic SERP accordion.
Why Did Google Deprecate the Rich Result in the First Place?
Google has not published a single definitive reason, but the pattern across its recent SERP feature decisions points toward a consistent theme: features that displaced organic click-through by answering the query directly in the results page, without adding proportionate value, have been steadily curtailed as Google's own AI Overviews increasingly take over that same direct-answer function natively. A visible FAQ accordion competing for the same attention and click-avoidance behaviour as an AI Overview creates redundant, potentially conflicting SERP real estate. Removing the older, more limited feature in favour of the newer, AI-generated one is a plausible and consistent explanation, even without an explicit statement from Google confirming it as the reason.
Should You Remove FAQPage Schema From Your Site Now?
No, for three separate reasons that hold independently of each other. First, the schema continues to provide the same citation-relevant structural signal to every AI engine other than Google's own generative features, and the independent research supporting that signal predates and is unaffected by this specific deprecation. Second, removing valid, accurate schema that matches your visible page content carries no benefit and a small but real cost: Google's crawlers still parse it, and a well-implemented FAQPage block continues to help disambiguate your content's structure for any system reading it, including Google's own systems during standard indexing, even without the old rich-result payoff. Third, the effort to maintain FAQPage schema on a page that already has genuine Q&A content is minimal once implemented — there is no meaningful engineering cost to leaving it in place, and removing it would require active work for no offsetting gain.
The one thing worth reconsidering is whether FAQPage schema should be treated as a priority engineering task specifically because of an expected Google rich-result payoff. That specific justification is now gone. If your organisation was implementing FAQPage schema primarily as an SEO team initiative aimed at winning the classic FAQ snippet, that particular business case no longer exists, and resourcing decisions should reflect the schema's actual remaining value — citation structure for non-Google AI engines and general content disambiguation — rather than the outdated SERP-feature justification.
What Should Change About How You Write and Prioritise FAQPage Content?
Three adjustments follow logically from the deprecation, each grounded in what actually still matters rather than what used to matter.
Stop optimising FAQ answer length for the old rich-result display constraints. The classic FAQ snippet had implicit display-length considerations tied to how much text Google would show in the expanded accordion. With that feature gone, the correct length guidance is now purely the AI-citation-oriented recommendation covered in the schema types guide in this series: 40 to 60 words per answer, direct and specific, optimised for extraction as a standalone citable passage by a non-Google AI engine rather than for legibility inside a now-nonexistent SERP widget.
Prioritise FAQPage implementation by non-Google citation opportunity, not by classic SERP visibility potential. Before the deprecation, a reasonable prioritisation heuristic was "which pages have the highest search volume for question-format queries where a rich result might realistically appear." That heuristic is now obsolete. The correct replacement, using the prompt-tracking discipline covered in the prompt engineering guide, is prioritising FAQPage implementation on pages that answer the specific questions your tracked prompts show competitors currently winning on Perplexity, ChatGPT, and Claude.
Do not conflate FAQPage schema with FAQ-style content generally. The deprecation affects the machine-readable schema markup's classic rich-result behaviour. It says nothing about whether writing genuine, well-structured FAQ sections into your pages is still good practice for human readers or for AI extraction generally. A well-written FAQ section without any schema markup at all still provides readable, extractable Q&A content that AI engines can parse from the visible HTML — the schema is an enhancement layer on top of that content, not a replacement for writing it well in the first place.
How Do You Audit Whether Existing FAQPage Schema Is Still Correctly Configured?
The deprecation is a reasonable trigger to run the standard schema health check covered in the schema errors guide in this series, since a meaningful number of sites implemented FAQPage schema years ago and have not revisited it since. Confirm the schema's questions and answers still match the visible page content exactly — a schema-content mismatch is a trust signal problem for AI engines regardless of whether Google's rich result still exists. Confirm the schema is present in the server-rendered HTML response rather than only in client-side JavaScript, using the same curl-based verification method covered throughout this series. And confirm each answer still falls in the 40-to-60-word range that AI extraction favours, rather than the potentially different length constraints that may have been optimised for the now-retired rich-result display.
How NotionCue Helps You Confirm FAQPage Schema Is Still Earning Citations Post-Deprecation
The most direct way to know whether your continued FAQPage schema investment is paying off after this deprecation is to observe actual citation behaviour on the engines where it still matters, rather than relying on assumptions carried over from the old SERP-feature era. The NotionCue Citation Tracker monitors your tracked prompts across ChatGPT, Perplexity, Claude, Gemini, and Google's AI surfaces on a weekly basis, and the response text it captures will directly show you whether the specific Q&A content in your FAQPage schema is being extracted and quoted by any of the non-Google engines — the concrete evidence that this schema investment continues to earn its keep even without the retired Google rich result behind it.
The NotionCue AI Crawler Audit confirms the more basic prerequisite this whole discussion assumes: that your FAQPage schema is actually reaching AI crawlers in the first place, in the server-rendered HTML response rather than trapped in client-side JavaScript. Since the classic rich-result feature's removal takes away one of the ways teams used to informally verify their schema was "working" — you could previously just check whether the accordion appeared in search results — losing that visible feedback loop makes a direct crawler-access check more valuable than before, not less.
Start your free NotionCue trial and confirm, directly in your own tracked data, that your FAQPage schema is still earning citations on the engines that continue to read it, rather than relying on the now-unavailable visual confirmation the old Google rich result used to provide.
If your organisation had built internal reporting or dashboards that tracked FAQ rich-result appearance as a KPI, that metric will now permanently read zero for every property, everywhere, and should be formally retired from any reporting deck rather than left in place generating a confusing flatline. Replace it with citation-rate tracking on the engines where FAQPage schema still has a documented, measurable effect.
Frequently Asked Questions About the FAQ Rich Result Deprecation
Did the deprecation happen gradually, or all at once?
The May 7, 2026 change was the second and final phase of a process that began with a 2023 restriction limiting the feature to a narrow set of government and health sites. The May 2026 deprecation notice removed it entirely, for all site types, all query types, and all languages, with no remaining exceptions.
Does this deprecation affect HowTo rich results in the same way?
HowTo rich results followed a broadly similar trajectory, restricted from most standard search results earlier in 2026, though the specific mechanics and timeline differ from the FAQ feature's history. The same underlying logic applies to both: the classic SERP-widget behaviour has been curtailed, while the schema type itself remains valid and continues to be read by AI systems during citation selection, as covered in the dedicated HowTo schema guide in this series.
Will Google eventually deprecate other schema-driven rich results the same way?
There is no official announcement of further deprecations at this time, but the pattern with both FAQ and HowTo rich results suggests Google is actively re-evaluating which classic SERP features still make sense alongside its own generative AI answer surfaces. Monitor Google's Search Central "What's new" changelog periodically for deprecation notices rather than assuming any specific rich result is permanent, and treat schema implementation as valuable primarily for its AI-citation function going forward rather than for any specific classic rich-result payoff.