Quick read: Google removed FAQ rich results from search on May 7, 2026 — the dropdown accordions that appeared below some search results are gone for most sites. An Ahrefs study tracking 1,885 pages found that adding JSON-LD schema produced no statistically meaningful increase in AI Overview, AI Mode, or ChatGPT citations. AI systems read visible HTML during retrieval, not JSON-LD markup. FAQ schema is still worth keeping if it's already on your pages, but the case for adding it has shifted: the value was always in the Q&A content structure, not the markup wrapper.
On May 7, 2026, Google updated its FAQ structured data documentation with a note at the top: FAQ rich results were no longer appearing in Google Search. Search Engine Land confirmed the full timeline: rich results gone in May, the Search Console FAQ report and Rich Results Test removed in June, API support ending in August.
If you noticed something missing from your Search Console reports, that's why.
So now comes the question everyone's website vendor is either quietly avoiding or loudly oversimplifying: is FAQ schema still worth having on your site? The answer is yes — but not for the reason most people are selling you.
What the rich result actually was
For about five years, the FAQ rich result was one of the most-chased features in small business Search Engine Optimization (SEO). Add FAQPage JSON-LD to a page, match it to visible Q&A content, and Google would sometimes show your questions expanded directly in the search result — taking up more visual space, looking more authoritative, occasionally pushing a competing result below the fold.
Two caveats made this murkier than it appeared. First, the feature was already being heavily restricted before May 2026 — Google had limited FAQ rich results primarily to government and health sites back in 2023. Most small business sites adding FAQ schema after that point weren't seeing accordion dropdowns; they were doing technical work that no longer produced a visible payoff. Second, even when it worked, it worked inconsistently. The feature showed up on some queries, on some days, for some sites. It was never reliable enough to build a strategy around.
May 2026 made the state of things official. The feature is gone. Whatever scraps of accordion remained no longer appear.
The Ahrefs study, and what it means for AI citations
The question underneath this one — the reason anyone still cares about FAQ schema now that the rich result is gone — is whether FAQ markup helps you get cited by AI systems. Google AI Overviews, ChatGPT, Perplexity, AI Mode. The pitch has been: add FAQPage schema, show up in AI answers.
There's data on this now, and it's worth reading directly before trusting anyone who summarizes it conveniently. Ahrefs tracked 1,885 pages that added JSON-LD schema between August 2025 and March 2026, matched them against 4,000 control pages without schema additions, and measured citation changes across Google AI Overviews, AI Mode, and ChatGPT.
The result: no meaningful uplift. Google AI Mode was up 2.4%; ChatGPT was up 2.2% — both within statistical noise. Google AI Overviews was actually down 4.6%, which was the only figure to clear the bar for statistical significance, and it went the wrong direction.
The explanation matters: during direct retrieval, AI systems don't read JSON-LD. They read visible HTML — the actual text on the page that any reader could see. JSON-LD is in the <head> or in a script block; the retrieval process ignores it entirely.
This is why the correlation that generated so much schema enthusiasm — AI-cited pages are roughly three times more likely to have JSON-LD than non-cited pages — is a correlation, not a cause. Sites that invest in structured markup also tend to be technically cleaner, better maintained, more authoritative, and more thoughtfully written. The schema travels alongside quality; it doesn't produce it. You can't schema your way out of a thin page.
What still works
Here's the part that didn't change: the Q&A format itself.
AI systems are in the business of fielding questions and returning answers. A page organized around specific questions — with direct, visible answers right below them — is exactly the format that gets extracted. Not because the page has a JSON-LD block, but because the content already reads like what the AI is trying to produce.
That's the actual argument for keeping FAQ sections on your pages: you're pre-formatting your content as the thing AI systems want to cite. The schema just describes what's already there. If what's already there is a real question and a real answer written clearly, the markup is metadata. If what's already there is vague or thin, the markup is a lie.
This connects to the broader work of Answer Engine Optimization — structuring your site so AI systems can extract, trust, and cite what you've written. The complete AEO guide covers this in depth. The short version: clear structure and direct answers outperform clever markup every time.
So should you keep FAQ schema or remove it?
If you already have FAQ schema on your pages: leave it. Google's documentation explicitly says you can keep it; it won't cause problems. Other search engines may still use it. The cost of having it is zero.
If you're building new pages: write real Q&A sections with clear question headings and direct answers, and let your SEO plugin handle the markup. Yoast, RankMath, and similar tools generate FAQPage JSON-LD automatically from FAQ blocks. You probably already have it without thinking about it.
If you're considering a focused schema project: know what you're buying. FAQ markup won't move AI citations on its own. For structured data that genuinely affects how Google represents your business — in Maps, in the Knowledge Panel, in local results — LocalBusiness schema is a separate type that does different, more direct work. That's the schema I'd prioritize on a new site before any FAQ markup.
What to spend the actual time on: writing the questions your customers ask on the phone, five times a week, and putting the answers on your pages where Google and ChatGPT can find them. Not keyword exercises. Real questions. "How long does this take?" "What's included?" "What happens if I need to cancel?" These live in a well-maintained FAQ section — and they work exactly as well as they did before May 2026, because their value was never in the accordion dropdown.
I built DBO's own blog with ### question format FAQ sections on every post, and that structure is what the site's extractFaqs() parser uses to auto-generate FAQPage JSON-LD. Not the other way around — the Q&A content came first, and the markup describes it.
What won't help
Adding FAQ schema to pages without real Q&A content. The Ahrefs data settles this. Schema doesn't substitute for content structure. If the questions aren't answered visibly on the page, the markup is inaccurate — and Google has flagged exactly this in its documentation.
Adding FAQ schema to chase AI citations. The study is clear: it doesn't move AI citations in any statistically meaningful way. If the pitch you're getting includes "add FAQ schema for AI visibility," the person making it has either not read the data or is hoping you won't.
Removing FAQ schema in a panic. Also not a productive use of time. Google said you can leave it. Leave it.
Confusing FAQ schema with other structured data types. FAQPage markup and LocalBusiness markup serve completely different purposes. FAQ schema describes Q&A content on a page; LocalBusiness schema tells Google who and where your business is. If you haven't set up LocalBusiness and Organization schema yet, that's the higher-leverage item.
FAQ
Is FAQ schema still worth using in 2026?
The markup itself doesn't move AI citations — Ahrefs tracked 1,885 pages and found no statistically meaningful increase across AI Overviews, AI Mode, or ChatGPT. But the underlying Q&A format remains valuable: AI systems extract structured question-answer content from visible HTML, so writing real FAQ sections in a clear Q&A format is still one of the most effective content structures for AI visibility.
When did Google remove FAQ rich results?
Google removed FAQ rich results from search on May 7, 2026. The Search Console FAQ report and Rich Results Test were removed in June 2026, with API support ending in August 2026. The change was confirmed in an announcement covered by Search Engine Land.
Should I remove FAQ schema from my site now?
No. Google's own documentation says the markup can stay — it won't cause problems and other search engines may still use it. There's no benefit to removing existing FAQ schema.
Does FAQ schema help with Google AI Overviews or ChatGPT?
Not on its own. The Ahrefs study — 1,885 pages with JSON-LD schema tracked against 4,000 controls — found no statistically significant improvement in AI citations on any major platform. AI systems read visible HTML during retrieval, not JSON-LD markup.
How is FAQ schema different from LocalBusiness schema?
FAQ schema (FAQPage type) marks up question-and-answer pairs on a specific page. LocalBusiness schema marks up your business's name, address, phone, hours, and service area, and influences how Google represents you in Maps and the Knowledge Panel. They're separate tools doing different work. LocalBusiness schema setup covers the latter type in detail.
What structured data should I actually prioritize?
LocalBusiness and Organization schema — the markup that tells Google who you are, where you operate, and what you do. These have more direct impact on local search representation than FAQPage schema. Article schema on blog posts, which reinforces authorship and publication dates, is also worth having. FAQ schema is fine to include but shouldn't be first on the list.
What actually gets a page cited in AI Overviews or ChatGPT?
Clear structure and direct answers. A page that opens with a specific question and answers it in the first paragraph — before expanding, qualifying, or elaborating — is what AI systems are built to extract. Question-format headings (H2s phrased as questions), visible Q&A sections, and content that doesn't make the reader work to find the point. Schema helps search engines understand what's there. But what's there is the thing that matters.
The May 2026 FAQ rich result deprecation is a clarifying moment, not a crisis. It removes one SERP feature that was already largely unavailable for most small businesses and confirms what the data now supports: structured markup describes quality, it doesn't create it.
Write the questions your customers actually have. Answer them directly and visibly. Let the schema describe that content accurately. That's the practice that transfers — to AI Overviews, to chatbots, and to whatever search infrastructure comes next.
If you want help getting the structured data, content strategy, and technical setup right from the start, the Get Found on Google and AI path covers how we approach it as a managed operation.
Want a back office for your site?
Builds and Care are by application. Quality projects only — quoted at application.
Apply for a build slot →RELATED READS