Google Just Killed FAQ Rich Results. Here’s What That Actually Means for Your Site.
The feature has been dying since 2023. May 7th was just the funeral. The interesting question is what the schema markup is actually worth now.
On Wednesday, May 7th, Google officially ended support for FAQ rich results in search. The expandable Q&A dropdowns that used to fan out under your search listing? Gone. The Search Console report that tracked them? Disappearing in June. The Search Console API support for the field? Gone in August.
If you missed the announcement, that’s because Google didn’t make one. They updated a developer documentation page and let the news do its own walking. Search Engine Land caught it. The SEO press picked it up by Thursday. Most business owners with FAQ schema sitting on their site still have no idea anything happened.
Here’s the part that’s missing from most of the coverage: this isn’t a sudden change. It’s the obituary.
A slow-motion funeral that started in 2023
Google restricted FAQ rich results to government and health sites in August 2023. For everyone else, the dropdowns disappeared from search results almost three years ago. If your dentist client or your plumber client has FAQ schema on their site and you check the live SERPs, you have not been seeing those rich results since the summer of 2023. They’ve been dead for a while. May 7th just made it official for the last few authoritative sites that still had them.
That timeline matters because it tells you something about how Google operates. They retire features quietly, in stages, on a documentation page nobody reads, and then act surprised when site owners are still implementing things the feature is no longer doing. The HowTo rich result went the same way in 2023 and 2024. The pattern is clear: anything that competes with AI Overviews for screen real estate gets phased out.
Which leads to the question every business owner should actually be asking: if FAQ rich results are gone, is the FAQ schema markup on my site doing anything at all anymore?
The vendor answer is mostly wrong
If you’ve searched for what to do about this, you’ve probably already seen the talking point: “FAQ schema is dead for rich results, but AI loves it now! Keep your FAQ schema for ChatGPT and AI Overviews citations!” That story is everywhere. It’s also mostly wrong.
Here’s what the actual evidence says.
Google’s own documentation states that no special schema markup is needed for AI Overviews or AI Mode. Direct quote from the team that ships the feature. Not me reading tea leaves. Google saying it on their own developer site.
A December 2024 study by Quoleady and Search Atlas found no correlation between schema coverage and AI citation rates. Sites with heavy schema were not cited more by AI search engines. Sites with minimal schema were not cited less. The variable that predicted citation was content authority and relevance.
A February 2026 test by Mark Williams-Cook showed that LLMs tokenize JSON-LD but do not parse it as structured data the way Google does. Translation: ChatGPT, Perplexity, and Claude can technically see the schema in your page source, but they’re extracting answers from your visible question-and-answer text, not from the JSON-LD payload.
So the honest version of the FAQ-schema-and-AI story is this: the markup itself is a weak direct signal. What gets you cited is the visible Q&A content on the page, written clearly, with answers in the 40-to-80-word range that AI systems can extract cleanly. The schema is a wrapper. The content inside it is what does the work.
There is one indirect pathway where the schema still earns its keep, and it’s worth understanding. Schema markup helps Google build out its Knowledge Graph representation of your site, which helps your organic rankings, and roughly 76% of AI Overview citations come from pages already ranking in the top 10. So FAQ schema can influence AI Overview citations, just not the way the vendor blogs say. It works through the side door, not the front.
What to actually do this week
If you have FAQ schema on your site, here’s the decision tree.
Leave it in place if:
- the schema accurately reflects visible Q&A content on the page (the answers in the markup match what a reader sees)
- the page itself is providing useful answers to real questions your audience asks
- you have no specific reason to clean up the codebase right now
There’s no penalty from Google for leaving valid schema in place. It does no harm. The Knowledge Graph pathway gives it a small but real upside. The cost of leaving it alone is zero.
Strip it out if:
- the schema is from an old plugin or template and doesn’t match the visible content (mismatched markup is a quality issue Google has been quietly penalizing for years)
- you have FAQ schema on pages where there are no actual visible FAQs (this used to be a common dirty trick to grab the rich result; Google’s ignored it for a long time and it’s pure technical debt now)
- you’re cleaning up a site and want fewer moving parts to maintain
Either way, here’s where the real work is:
Stop thinking about FAQ schema as a feature you check off and start thinking about FAQ content as something AI systems actually extract. The markup is incidental. The content is the asset. That means writing real answers to real questions, in plain English, in chunks an AI can lift cleanly. Forty to eighty words per answer. Direct. No throat-clearing. No “Great question!” openers. The answer comes first, the context comes after, and the format mirrors how someone would actually ask the question out loud.
If your FAQ section reads like a sales objection-handling document with the word “affordable” in every other answer, AI systems will skip you and cite a competitor whose answers actually answer the question. That has been true for a while. May 7th just made it more obvious.
The bigger pattern this fits into
Strip away the specifics and this is a simple story. Google retired a feature that gave websites visible credit on the search results page, in favor of features that absorb the answer and keep the user on Google. FAQ rich results sent clicks to your site. AI Overviews don’t. The math for Google works. The math for the small business that depended on those clicks doesn’t.
That’s not a reason to panic, and it’s not a reason to chase whatever new schema type the next vendor blog post tells you to implement. It’s a reason to stop optimizing for visual real estate on a search results page and start optimizing for being the source AI systems extract from. Different game. Same fundamentals: clear writing, real expertise, content that answers the question someone actually asked.
Field Notes are where I write down the things I run into that are worth telling other people about. This one’s short and direct because the news itself is short and direct. Google killed a feature that’s been mostly dead for years. The schema markup behind it still has a small role to play, but not the one most blogs are claiming. Write your FAQs like a person answering a question, leave or remove the markup based on whether it matches the page, and spend the time you saved on something that actually moves your business.
Sources for this Field Note: Google Search Central FAQPage documentation (May 8, 2026 update); Search Engine Land coverage of the May 7 deprecation; the Quoleady/Search Atlas schema-vs-citation study (December 2024); Mark Williams-Cook’s February 2026 LLM tokenization test.
