Google Was Bound to Kill-off the FAQ Rich Result
On May 7, 2026, Google officially killed-off FAQ rich results. However, it wasn’t a shocking move since as of August 2023, FAQ rich snippets were only available for government and health related websites. As of April 2026, the update also had nearly zero impressions for most users.
“FAQ rich results were bound to be do away with for quite some time and this is just a cosmetic change.”
— Abhishek Iyer, founder of ACME.BOT
What happened?
- FAQ rich results SERP treatment: Gone.
- The schema itself: Still parses. Google redirects to Article, QAPage, or DiscussionForum instead.
- The real shift: AI search made FAQ markup redundant.
It’s not about the deprecation. AI search is already using the same info FAQ structured data is designed to provide. It’s not about losing visibility because you lost a checkbox. It’s about losing visibility because search changed. The real question: what structured data is AI search still rewarding?
What Google Removed (and What It Didn’t)
It was a surgical removal. First, they removed the FAQ rich result display in search results, the FAQ report in Search Console, and FAQ markup’s support in the Rich Results Test in June 2026. This was followed by removal of FAQ in Search Console API in August 2026.
The FAQPage schema? It’s still valid. Although, it might not pull up as a rich result in Google search. Instead, Google has started to recommend publishers to use QAPage, Article, or DiscussionForum markup for similar content.
| Gone | Still Works |
|---|---|
| FAQ rich result SERP display | FAQPage schema parsing |
| Search Console FAQ report | Schema in general (redirected to alternatives) |
| Rich Results Test support | JSON-LD markup validation |
| Search Console API support | Structured data on your site |
With the exception of authoritative government and health publishers, who had exclusive eligibility for FAQ rich results as of August 2023, this change will have an almost negligible effect on most commercial sites.
Here’s what needs to be considered: By March 2025, 58% of users encountered an AI Overview when searching, and when one was present, clicked on traditional links around only half as often. The big change isn’t that Google won’t display the schema but that AI search may start stealing attention to your link or anchor text before they even get there.
The schema only seems to matter as an afterthought once AI Overviews were identified as the answer by Google.
AI Search and Schema Markup: Don’t Overrate the Checkbox
Will adding Schema markup really get your page more AI citations?
Not really. Google’s own guide on optimizing sites for AI note that structured data isn’t required for generative AI search. Structured data also signals to Google that a page is eligible for rich results in traditional search. But that’s about it.
Why has this perspective been on the mind of AEO practitioners for so long?
Sure, this is an easy box to tick. But, as Abhishek Iyer, founder of ACME.BOT, points out, is both understandable and overstated.
“Reality is that AI is extremely good at extracting structure and doesn’t necessarily prioritize the pages that markup.”
— Abhishek Iyer, founder of ACME.BOT
So, what do independent tests actually show?
Tests that pulled “direct-fetch” content found that AI systems would not pick up JSON-LD schema if the content was also not visible in the text on the page itself. While some tests claim that there’s a lift in citations — like 3x as likely, or 44% higher with schema, neither divulge their methodologies.
Takeaways:
- Schema is like metadata that’s invisible to users and LLMs that parse visible content
- When running searches to directly fetch schema, no structured data that was hidden was extracted from the page
- It isn’t clear how citation lift claims were calculated
The bottom line: Schema is a form of metadata. But AI doesn’t actually need metadata – it reads your page. It’s not the lever people think it is.
What Actually Affects Visibility in AI Search
An AI program isn’t interested in your markup. Instead, it’s looking at your words to determine if it can pull a complete answer from them.
Here are three things that really matter:
- Problem-solution paragraphs that are self-contained. The problem and answer should exist in the same paragraph and be very clear and complete, with no jumping around to other sections/context. While Google doesn’t really care about your JSON-LD or AI structure tags, it does extract from the structure of your prose.
- Scannable, descriptive structure. Well-organized pages with clear H2s and H3s are extracted by AI answers more often, according to Jason Pittock’s 2025 research.
- Queries go from fan-out and long-tail coverage. When you search on answer engines, they tend to spend most of their search budget on long-tail queries. So instead of one query like “best coffee makers,” you might search “best coffee makers under $100” or even “best coffee makers for small apartments.” That’s breadth that AI systems actually reward.
“Focus on getting their own voice heard as opposed to having an ‘AI voice.’”
— Abhishek Iyer, Founder, ACME.BOT
Your page is kind of like a self-service answer booth. If a reader who knows nothing about your topic reads a section and walks away with a solid answer, the AI can too. That’s the real lever.
Schema Types You’ll Actually Want to Keep in 2026
While a few schemas got the axe, these are the ones that really help you see a boost in search visibility.
- Article schema — Ensures your authorship is shown and that you stay eligible for news content and blog rich results. If you publish content regularly, this one’s non-negotiable.
- Product schema — Availability, ratings and pricing. In particular, e-commerce websites should consider implementing this. Data shows it provides the most amount of lift to CTRs of any schema type.
- BreadcrumbList — Lets AI systems look for context and still keep your architecture legible. It also props up your internal linking strategy in the eyes of the crawlers.
- QAPage and DiscussionForum — Explicitly broadened by Google as alternatives for FAQ rich results. These also serve as your next task if you have FAQ markup previously.
According to Digital Applied, Google only recommends JSON-LD — not RDFa or Microdata. These two forms are harder to manage and more prone to errors. For 2026, you may want to base your strategy solidly on eligibility for rich results instead of the variance of chasing AI citations.
Stop Chasing Markup, Start Building Readable Structure
The removal of FAQ rich results changed the way the search results page looks, but didn’t change how search engines fundamentally work. Schema markup remains important for eligibility of traditional rich results, but Google has directly confirmed you don’t get the AI citations for it. What remains is clearly structured prose, topical depth that covers long-tail angles and self-contained sections written in a distinctively ‘human’ voice.
“The truth of AI is that while it’s extremely skilled at extracting structure, it isn’t necessarily focused on the a page’s markup”
— Abhishek Iyer, founder of ACME.BOT
This is something that ACME.BOT, which sees that humans voice and structured, readable writing is your main optimization lever, not schema checklists, applies within its content workflows. From a practical perspective, review your existing FAQ pages for quality; move any Q&A content that fits into QAPage or DiscussionForum markup; then, switch up more of your writing towards creating self-contained answer paragraphs. That’s the play.