FAQPage schema and AI answers: does structured data still matter in 2026

No. Google removed the FAQ rich result from Search on May 7, 2026, so FAQPage schema no longer earns a snippet there. It still has a job: giving AI answer engines a clean, structured Q&A to read. That's a separate question from whether Google shows it on the results page.

What FAQPage schema is and why it mattered

FAQPage is a schema.org type: a WebPage presenting one or more frequently asked questions, each with its accepted answer. For years, marking a page up this way could earn an expandable FAQ snippet directly in Google's search results, extra vertical space that pushed a listing further down the page before a user even had to click.

That incentive was already narrowing before it disappeared. In a September 2023 documentation update, Google restricted the FAQ rich result to well-known, authoritative government and health sites, after years of unrelated sites bolting FAQ markup onto any page to claim the extra space. The pattern was predictable: a tactic that works gets copied onto pages it was never meant for, the search results page gets cluttered with low-value FAQ blocks, and the platform tightens eligibility until the practical answer for most site owners is that the feature simply isn't for them anymore.

Then, in 2026, Google finished the job. The FAQ rich result stopped appearing in Search on May 7, 2026. The documentation page describing the feature was removed in a June 2026 changelog update, and Search Console API and report support for it is scheduled for removal in August 2026. Existing FAQPage markup on a site isn't penalized, it just stops doing the one thing most site owners added it for. That's why it's fair to ask, in 2026, whether the feature is still worth the engineering time.

If you want the fuller picture of what schema markup can and can't do for a site in 2026, our broader schema markup guide to rich results covers the other types that are still live.

What it is and why it matters now

The schema.org definition of FAQPage hasn't changed. It still sits at WebPage > FAQPage in the type hierarchy, and it's still widely used: an estimated 1M to 10M domains carry FAQPage markup as of August 2026, based on Google's own web index aggregation. That's a lot of sites keeping code that no longer buys them a search snippet.

The reason is straightforward. A well-written FAQ section, wrapped in FAQPage markup or not, gives a reader (or a language model reading on a reader's behalf) a fast, unambiguous answer to a specific question, plus a clear boundary around what the page's owner does and doesn't claim. That's exactly the shape of input an AI answer engine wants when deciding what to quote back to a user. It's the core idea behind answer engine optimization: structuring content so a machine can extract a correct, attributable answer, independent of whether a traditional search engine rewards the structure with a visual snippet.

The FAQPage schema itself isn't doing that extraction work directly, though. The markup is metadata that tells a crawler "this block is a question and this block is its answer." What actually gets quoted is the plain-text answer, whether or not a script tag wraps it. That distinction matters for anyone deciding where to spend limited engineering time: the markup is a small, low-cost addition, but the underlying writing is where the real gain sits.

How it works in practice

The open question is whether the markup itself, separate from the quality of the writing underneath it, moves the needle on being cited by an AI system. A 2026 Ahrefs study tried to isolate that. Researchers tracked 1,885 pages that added JSON-LD schema, including FAQ, Article, Product and HowTo types, between August 2025 and March 2026, and compared them against 4,000 control pages that added none.

The result was a flat line. Google AI Overview citations for the schema pages fell 4.6%, a small but statistically significant decline. Google AI Mode citations rose 2.4% and ChatGPT citations rose 2.2%, both changes small enough to be indistinguishable from random noise. The study's own conclusion: adding schema produced no major uplift in citations on any platform tested.

One limitation is worth carrying forward before you draw a conclusion of your own. The sample was pages already receiving 100 or more AI Overview citations before schema was added. So the study measures what happens when you add markup to a page an AI system has already found and trusts, not what happens on a page no AI system has indexed yet. Call it evidence against schema as a discovery shortcut, not proof that structure is irrelevant everywhere. Our guide on getting cited by ChatGPT, Perplexity and Google AI Overviews goes into what does correlate with citation rate, and markup isn't the lead variable.

Tradeoffs and edge cases

There's still a case for FAQPage markup, just a narrower one than it used to be. It's worth adding to a page that already has a genuine, specific set of frequently asked questions, the kind visitors actually type into a search box or ask a chatbot. On those pages, the structured Q&A format forces a discipline that plain prose doesn't: one clear question, one direct answer, no hedging. That discipline is what makes content easy to quote, and our note on answer capsules AI assistants actually quote covers that pattern in more detail.

It isn't worth adding to a page just to claim a rich result, because that result no longer exists. And it isn't worth padding an unrelated page with a generic FAQ block purely for the markup. When we scope a structured data pass for a client at Kallos Labs, the first question is always whether the FAQ content would earn its place on the page even with no schema attached. If the answer is no, the markup isn't going to fix that.

Here's a simple way to decide: list the three to five questions a real visitor asks before buying, signing up, or booking a call with your team. If those questions already exist somewhere on the page in prose form, formatting them as an explicit FAQ section with FAQPage markup costs little and gives an AI system a cleaner target to quote. If you find yourself inventing questions nobody asks just to fill out a section, that's the sign the FAQ is decoration, not content, and no amount of markup changes that.

The practical takeaway for 2026: nothing needs to replace FAQ schema as a tactic, because it was never a growth lever on its own. The work that matters is the same work it always was, writing FAQ content specific enough to be genuinely useful, whether or not Google decides to render it as a snippet.

Frequently asked questions

Did Google remove FAQ rich results from search?

Yes. Google stopped showing the FAQ rich result in Search on May 7, 2026, removed the feature's documentation in June 2026, and is removing Search Console API and report support for it in August 2026.

Should I remove FAQPage schema from my site now that Google dropped the rich result?

No, not automatically. The markup no longer earns a Search snippet, but it still gives AI answer engines a clean, machine-readable Q&A structure to parse, and removing it gains nothing since the markup itself carries no penalty.

Does adding FAQ schema get you cited more often by ChatGPT or Google AI Overviews?

Not reliably on its own. A 2026 Ahrefs study of 1,885 pages found that adding JSON-LD schema, FAQ included, produced no major uplift in AI citations on any platform tested.

What should replace FAQ schema as an SEO tactic in 2026?

Nothing needs to replace it, because it was never a growth lever by itself. The actual work is writing genuinely useful, specific Q&A content, since that is what both readers and AI systems reward, with or without a rich result attached.

If you're deciding how much of your structured data strategy to rebuild around AI answer engines rather than classic Search snippets, our SEO and AEO work is where we help clients make that call.