Daniel Carter
Quick housekeeping, the title is technically wrong.
Google isn't killing FAQ schema. Google is killing FAQ rich results - FAQ structured data still has a place in society, just less so in SEO.
There's a difference and it matters because a lot of SEOs are about to either panic-strip JSON-LD off their sites or pretend nothing happened, but, in reality, there's no need to do anything, you won't be penalised for leaving structured data in place, other sources may use it still.
I've been doing SEO since 1998 and FAQ schema is the single biggest example I can think of where the SEO industry took a perfectly good Google feature and abused it into oblivion, well that and the old "aggregateRating" which a huge volume of sites abused - every page would have gold star review ratings appended to it (until Google wiped that out).
It got what it deserved.
Here's what happened on May 7, 2026, what you could do this week, and why this is a slightly different conversation to the one most of the noise on LinkedIn is having.
What Google announced on May 7, 2026
Google updated the official FAQPage structured data documentation on May 7, 2026 ( https://developers.google.com/search/docs/appearance/structured-data/faqpage)with a deprecation notice at the top of the page. The notice is worth quoting directly from Google's developer docs:
"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. To allow time for adjusting your API calls, support for the FAQ rich result in the Search Console API will be removed in August 2026."
Three things to take from that:
As of May 7, 2026 the FAQ rich result no longer renders in Google Search results. Gone. Today.
June 2026 - the Search Console search appearance report for FAQ goes, and the Rich Results Test stops validating FAQ schema as an enhancement.
August 2026 - the Search Console API stops returning FAQ rich result data. Anyone with a dashboard pulling that field needs to update it before August.
That's the timeline. The visible feature in the SERPs is dead and the supporting reporting infrastructure is being shut down over the next three months.
What Google did NOT say
This is where most of the takes fall over.
Google did not say: (and take note here!)
Remove your FAQ schema.
FAQ schema doesn't help your SEO.
FAQ content is no longer relevant.
What Google said is that the rich result is gone. Schema and rich results are two different things.
Rich results are what Google decides to display in the SERP. Schema is how Google understands what a page contains.
Google can sunset the visible feature without abandoning the underlying signal. That's exactly what's happening here. If you have FAQPage JSON-LD on your site that genuinely describes question-answer pairs that actually exist on the page, Google can still parse it and use it for comprehension. They've just stopped rendering it as expandable dropdowns in the SERP.
This isn't speculation. Google's documentation explicitly leaves the schema valid for site owners who want to keep it.
Why FAQ rich results got killed
I've audited a lot of sites over the last few years and I can tell you exactly why this happened.
Sites stuffed FAQ schema everywhere it didn't belong:
Product pages with 15 fake FAQs designed to occupy more SERP real estate
Service pages with question-answer pairs nobody had ever asked
Blog posts where the "FAQ" was scraped from People Also Ask on the same query
Programmatic templates pumping out the same 5 questions across thousands of pages
Pages where the answers were keyword-stuffed paragraphs designed to pad word count
Affiliate pages where the FAQ was three-quarters of the visible content
The August 2023 restriction that limited FAQ rich results to "well-known, authoritative government and health websites" was Google's first attempt to clean it up. It didn't fully solve the problem, because Google's reporting infrastructure still tracked the enhancement and a long tail of authority sites kept the feature.
The May 7 announcement closes the loop. The whole feature is gone for everyone, including the gov and health domains that retained it after 2023.
This is what happens when SEOs collectively abuse a feature. It gets removed. We did this to ourselves.
There's also the reality that Google's AI Overviews are becoming more and more prominent and, they're starting to blend it with Google's AI mode, this is, effectively undermiming the point of PAAS (PEOPLE ALSO ASKED) features - plus, a huge amount of PAA's over the years have been wrong, poorly written, spammed etc.
A lot of PAA's are generally "not that good or overly helpful"
Should you remove the schema?
Two groups of SEOs have already formed online around this
Group 1: Get rid of it, the feature is now dead, of no value, the JSON-LD is now just pointless code, get it off a site if it doesn't serve a purpose.
Group 2: leave it alone. Google still parses it, it doesn't hurt anything, why are you wasting dev time on cleanup?
My honest take? It depends entirely on whether your FAQ content is real and of some actual value to target users.
If your site has FAQPage schema describing questions and answers that:
Actually exist as visible content on the page
Are genuine user questions, not invented for SEO
Have answers that genuinely address the question
Then leave it alone. The cost of keeping valid schema on the page is pretty much zero. The schema still helps Google interpret the page structure, and there's a reasonable argument that AI retrieval systems read JSON-LD as part of their pipeline. More on that below.
If your site has FAQPage schema that:
Marks up content that isn't actually visible on the page
Was added programmatically to inflate SERP appearance
Contains questions nobody asked, written purely for keyword targeting
Is duplicated across hundreds or thousands of templated pages
Then you have a different problem. Bin the schema, but more importantly audit the FAQ content itself. If the markup was junk, the underlying content probably is too. Removing the schema and leaving spammy FAQ blocks live on the page doesn't fix anything. That content is the actual issue.
SearchPilot ran an A/B test in late 2024 on removing FAQ schema and found no statistically significant effect on organic traffic. So nobody is going to nuke their rankings by leaving it alone or removing it. This isn't a high-stakes decision. It's a hygiene one.
What you definitely shouldn't do is panic-strip JSON-LD from your site this week because you read a LinkedIn post about Google "killing schema". That's not what happened.
Does FAQ schema still help with AI Overviews and LLM citations?
This is the more interesting question and the answer isn't clean.
The optimistic position: AI search engines like Perplexity, ChatGPT, Gemini and Google AI Overviews can read structured data and use it to interpret pages. Therefore, keeping valid FAQ schema gives you a citation advantage.
The realistic position: most LLM-based retrieval systems work primarily off the rendered DOM. They tokenise visible text. JSON-LD lives in the page source but it doesn't add anything that a parser of clean HTML doesn't already capture. People Also Ask, the closest legacy analogue, has always extracted question-answer pairs from on-page semantic HTML, not from schema.
I lean toward the realistic position.
In my own testing on client sites, the strongest signal for getting an answer cited in an AI Overview or pulled into an LLM response is having the question in a clear H2 followed by a tight, complete answer in the next paragraph. Semantic HTML beats JSON-LD for extraction.
That doesn't mean schema is useless. It means schema is doing comprehension work, not extraction work. The two are different.
Practical translation: don't add FAQ schema specifically to chase AI citations. Write the question in an H2 and the answer underneath it, and let the visible content do the work.
What you could do this week
If you've got time on Friday afternoon:
Pull a list of pages on your site that have FAQPage structured data. Screaming Frog or Sitebulb will give you this in 10 minutes.
Export your historical FAQ rich result data from Search Console before the report goes in June 2026. You won't get it back once it's gone.
Compare CTR on those pages 28 days before and 28 days after May 7, 2026. This is your real impact assessment, not the LinkedIn takes.
For pages where the FAQ content is real and useful, leave the schema alone.
For pages where the FAQ was templated junk, audit the content not just the schema. Strip the spammy questions. Rewrite or delete the answers that exist purely to pad keywords.
Update your QA checklist. Stop using the Rich Results Test as the gate for FAQ schema validation from June 2026.
If you've got dashboards or pipelines pulling FAQ rich result data from the Search Console API, schedule the update for August 2026.
That's the entire to-do list. It's a hygiene exercise, not a five-figure project.
What this tells us about structured data going forward
I've been around long enough to remember when meta keywords mattered. I've watched Google sunset Ask, Hotbot, Geocities and fifty other things over 25 years. Features come and features go. Schema is no different.
A few things to take away:
Structured data types tied to specific visual SERP features are always exposed to deprecation risk. Google launches features, the SEO industry abuses them, Google removes the visual treatment. FAQ. How-to. COVID. JobTraining. Same story every time.
Schema types tied to durable user value tend to survive. Product, Article, Organization, Breadcrumb, Review. These aren't going anywhere because they describe real-world things that Google needs to understand to do its job.
The "structured data as a citation hack" framing is going to keep losing. The "structured data as comprehension and entity-building infrastructure" framing is going to keep winning. Build for the second one.
Don't build any SEO strategy around a single rich result. If your traffic was meaningfully dependent on FAQ dropdowns, that was always a fragile position.
John Mueller already addressed similar panic in late 2025 when an earlier round of schema deprecations was announced. His framing was clear: Google isn't killing schema, it's pruning the markup types that don't earn their keep. Some of the panic on LinkedIn this week is people who didn't read what was actually said.
Long story short
The FAQ rich result is dead at "googles" level, the FAQ schema is not.
If your FAQ markup describes real content that helps users, leave it. If it was a CTR scam, fix the underlying content first and worry about the schema second.
This isn't a crisis. It's a long-overdue cleanup of a feature that the SEO industry broke through abuse. The same pattern will play out again with the next rich result we collectively over-exploit.
Build SEO for the bits that don't get sunset. Real expertise, real content, structured data that earns its place, and a brand worth understanding.
Peace out.