Google killed FAQ rich results. Schema isn't dead — it changed jobs.

What Google actually said

Strip the panic and the change is narrow and specific. Google added a deprecation notice to its FAQ structured-data documentation, and the wording is worth reading exactly as written rather than as it was relayed:

"As of May 7, 2026, FAQ rich results are no longer appearing in Google Search. We will be dropping the FAQ search appearance, rich result report, and support in the Rich results test in June 2026 … support for the FAQ rich result in the Search Console API will be removed in August 2026."
— Google Search Central, FAQ structured-data documentation, May 2026

Three dates, three sunsets. The visual feature stopped rendering on 7 May 2026. The reporting — the FAQ search-appearance filter, the rich-result report, and Rich Results Test support — goes in June 2026. The Search Console API field for it goes in August 2026. That's the whole change. Note what it does not say: it never tells you to remove your markup.

Search Engine Journal's Matt G. Southern, reporting the change, put the crucial caveat plainly: FAQPage is still a valid Schema.org type, and Google "has previously said that unused structured data doesn't cause problems for Search." Nothing breaks if you leave it. The markup is valid; only the SERP feature it used to unlock is gone.

This is the third act, not a plot twist

The reason to keep calm is that we've watched this exact move three times now. Google has been quietly retiring rich-result features — not the underlying vocabulary — for years, on a clear logic: low usage, low value, or a feature that no longer earns its place on a busier results page.

HowTo rich results were cut from desktop in 2023. Then on 12 June 2025, Google retired seven structured-data types in one sweep — Book Actions, Course Info, Claim Review, Estimated Salary, Learning Video, Special Announcement, and Vehicle Listing (Search Engine Journal, June 2025). FAQ in 2026 is the same play. Each time, the SERP feature disappears, the Schema.org type stays valid, and rankings are explicitly unaffected. The vocabulary isn't being deprecated. The decoration is.

Read as a sequence, the message is unmistakable: the results page is shedding bolt-on visual enhancements. If your structured-data strategy was built to harvest those enhancements, the ground really is moving under you. If it was built for something else, this barely registers.

"Dead" and "AI superpower" are both wrong

The discourse collapsed into two confident, opposite mistakes. They're worth naming because acting on either one costs you.

Both camps make the same error: they treat structured data as a lever you pull for a reward Google hands out — a rich result, an AI citation. It was never that. Its real job is duller and more durable, and it's the same job it always had.

Entity legibility, not SERP real estate

Structured data is, and always was, a way to state unambiguously what the things on your page are. This is an Organization, with this name and these profiles. This is a Product, with this price and this availability. This page sits here in the site (BreadcrumbList). This is an Article, by this author, published on this date. This is a LocalBusiness, at this address, open these hours. That work — entity disambiguation — got more valuable the moment machines started reading pages to answer questions instead of just ranking links.

The FAQ feature was the part of schema that bought SERP real estate — pixels on the results page. That's the part Google reclaimed. What it left untouched is the part that makes your entities legible to extraction: clean, explicit statements a model can lift without guessing. The job didn't shrink. The flashy half got cut and the load-bearing half stayed.

There's one non-negotiable rule that the AI shift makes sharper than ever, and Google repeats it in its own myth-busting: your structured data must match the visible content of the page. Markup that claims a price the page doesn't show, or an FAQ that isn't on the page, isn't a clever hack — it's a mismatch that can earn a manual action, and worse, it can feed a model the wrong number to quote back to a customer. Schema is a description of what's there, not a backdoor to say something else.

And the headline framing to keep: Google's own line is that "GEO is still SEO at the core." The fundamentals that earn you a ranking — clear content, accurate markup, real authority — are the same ones that get you read and cited by AI features. There's no separate AI game with separate rules. There's the same game, now also played by readers who happen to be models.

A short, unpanicked checklist

None of this is urgent in the "site is on fire" sense. But there are a few concrete moves worth making before the June and August sunsets land.

Tools for the job schema actually has now

If schema's job is entity legibility, the work is to find where your markup is missing, wrong, or out of parity with the page — and fix it. That's what we build for.

FAQ rich results are gone, and that's fine. The reason to do structured data never lived in that drop-down. Schema's job is to make your entities unambiguous to whatever is reading the page — and in 2026, more and more of what's reading is a model deciding what to quote. That job didn't end on 7 May. It got more important.

Audit your structured data