If you marked up your pages with FAQ schema chasing those expandable question-and-answer blocks in the search results, that game is over. Google switched them off, and it did not even write a blog post about it.
On May 7, 2026 the company added a quiet note to the top of its FAQ structured data documentation: FAQ rich results no longer appear in Google Search. No webmaster announcement, no fanfare, just a label on a developer doc. The Search Console report and Rich Results Test support were to follow in June 2026, and the Search Console API support to end in August 2026. Both have since happened (update at the end).
A lot of people read this as "schema is dying." We read it as something more useful. A week later, on May 15, Google published a new AI search optimisation guide that, among other things, tells you not to bother with special schema for AI features. Put the two together and the message is hard to miss. The era of bolting on markup as a growth hack is closing. Here is what that actually changes for businesses in Cyprus and the US, and what to do instead.
What Google actually did
First, let us be precise, because the panic is bigger than the change.
FAQ rich results are the visual feature, the accordion of questions that used to expand under some search listings. That feature is gone for everyone. It is not "limited to health and government sites" anymore, the way it was after Google's 2023 cutback. It is switched off across the board.
The FAQPage schema type itself is not gone. It is still valid Schema.org markup. Google has been clear that leaving it on your pages causes no harm, throws no errors, triggers no manual action, and will not touch your rankings. As Search Engine Land summarised it, you "can remove the FAQ structured data from your code, if you want, but you can also leave it." Other search engines may still read it.
So nothing breaks if you do nothing. What broke is the reason most people added it in the first place: the visible reward in the SERP.
Why this matters more than it looks
Here is the part worth sitting with. For years, structured data was treated as a vending machine. Put markup in, get SERP decoration out. FAQ blocks, review stars, recipe cards. The whole pitch was extra real estate.
That framing was always a bit backwards, and FAQ rich results just proved it. The schema was never doing the work. Your content was. The markup was a wrapper that earned you a bigger slot in the results, and Google decided that slot was cluttering the page and pulled it.
This is the same lesson we have been hammering about llms.txt and the wave of AI-specific markup. When the only reason you are adding something to your site is to game a feature, you are exposed the moment the platform changes its mind. And platforms change their mind constantly.
Google more or less said the quiet part out loud
The FAQ change did not arrive in a vacuum. In its new AI search guide, Google went out of its way to knock down the idea that AI search needs special technical treatment. The relevant lines are blunt.
On AI-specific files: "You don't need to create new machine readable files, AI text files, markup, or Markdown to appear in Google Search." On schema: "Structured data isn't required for generative AI search, and there's no special schema.org markup you need to add." On chunking your content for machines: "There's no requirement to break your content into tiny pieces for AI to better understand it." We unpacked what Google actually said about chunking separately.
And the headline position, the one that settles a year of consultant arguments: "From Google Search's perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO." Search Engine Journal covered the guide in full, and the framing matches what we have argued in our own AEO breakdown and in our read on Google's "still SEO" line: the acronyms are new, the work is not.
So in the space of about a week Google removed a popular schema feature and told everyone to stop treating schema as a special lever for AI. That is a consistent position, not two random events.
Our take: do not confuse "schema for show" with "schema that means something"
We want to be careful here, because the lazy takeaway is "schema is dead, rip it all out," and that is wrong.
Structured data still earns rich results for the types Google actually supports: Product, Review, Breadcrumb, Local Business, Article, Event, and the rest. Those are not going anywhere. Mark up a product with a real price and availability and you can still earn the visual treatment, and you give Google's systems a clean, unambiguous read of your facts. That second part matters more every quarter.
The distinction we draw with clients is simple. There is schema that describes something true about your business that a machine would otherwise have to guess, your prices, your locations, your opening hours, your author credentials. Keep that, improve it, treat it as part of how you communicate facts to systems that read your page instead of looking at it. The way Google's AI features pull content through retrieval and query fan-out rewards pages where the facts are stated cleanly, whether or not there is markup around them.
Then there is schema added purely to trigger a SERP feature, with no other purpose. FAQ markup was mostly the second kind. That is the kind Google just taught us a lesson about.
What to actually do now
You do not need to do anything urgent. Nothing is on fire. But here is where we would spend the effort.
Leave your FAQ markup or remove it, your call, but stop counting on it. If it is already there, it does no harm. If you were about to spend a sprint adding it across a site to win rich results, cancel that. Spend the time on the content instead.
Keep the FAQ content, lose the schema mindset. Genuinely useful question-and-answer content still helps you, because it answers the long-tail and conversational questions people actually ask AI tools and search engines. The HOTH put it well: your FAQ content still matters, just not for the reason you thought. Write it for the human and the model reading it, not for an accordion that no longer exists. That makes it a content strategy question now, not a markup one.
Audit the structured data that still pays. Make sure your Product, Local Business, Article and Breadcrumb markup is accurate and complete, because those features are alive and those facts are exactly what AI systems lift. If you are rebuilding any of it, our JSON-LD schema markup generator writes valid Article, Product and Local Business blocks. If you are not sure what is on your pages or whether it is valid, that is the kind of gap a proper technical audit is built to surface.
Stop chasing AI-specific gimmicks. Google just told you llms.txt, special markup and content chunking are not requirements for its AI features. Believe it. The compounding work is helpful content, a clean crawlable site, and being a recognised source, not a folder of machine-readable files nobody asked for.
Update, September 2026
The wind-down finished on schedule. Google removed the FAQ rich result documentation on June 15, along with the Search Console rich result report and Rich Results Test support, and the API support ran out in August as announced (Google Search Central updates, FAQPage documentation). On the same June day Google added a line to its docs saying llms.txt files are not needed for Google Search and do not affect visibility or rankings, which closes the loop on the AI-specific-files argument above.
No further structured data types were deprecated between June and the start of September. The supported list in the section above still stands.
The bottom line
The death of FAQ rich results is a clarifying moment, not a tragedy.
For a decade, structured data got sold as a shortcut to SERP features. Google has now removed one of the most popular of those features and, in the same breath, told everyone that AI search does not need special schema either. The signal is consistent: markup is for describing what is true, not for decorating the results page. The sites that win are the ones whose facts are clean and current and whose content is genuinely the best answer, the same boring discipline that has always worked.
If you want to know which of your structured data is doing real work and which was just chasing a feature that no longer exists, a free SEO review is the quickest way to find out. We work with businesses across our home market of Cyprus and the US on exactly this, from technical groundwork to AI search optimisation, so your pages are legible to whatever is reading them next.
