Is FAQ schema still worth adding in 2026?

Yes, but for a different reason than before. FAQ schema no longer earns a visible rich result in normal Google search. However, FAQPage is still valid structured data that Google parses, and the question-and-answer format it encourages is one of the most extractable, citable shapes of content for AI assistants. Add it for AI readability and clean structure, not for a search-result badge.

Did Google remove FAQ rich results?

Yes, completely. In August 2023 Google restricted FAQ rich results, the expandable questions shown beneath a search listing, to authoritative government and health websites. It narrowed eligibility again in March 2026, and on 7 May 2026 it retired the feature outright for every site with no exceptions. No business today should expect FAQ markup to produce a visible enhancement in Google's results.

Does FAQ content help you get cited by AI like ChatGPT or Perplexity?

It genuinely helps, though mostly because of how it makes you write rather than because of the markup. AI assistants retrieve content at the passage level, lifting individual answers rather than reading a whole page. A question-shaped heading with a direct, self-contained answer underneath is exactly what those systems can extract and cite, and writing real FAQs enforces that structure.

What is the difference between FAQ rich results and FAQ structured data?

FAQ rich results were a visible Google search feature, the expandable questions displayed under a listing, retired completely on 7 May 2026. FAQ structured data, meaning FAQPage schema, is the invisible code describing the question-and-answer relationship to machines. Google removed the visible feature but still reads the structured data, and it remains useful to both Google and AI systems.

Why do people disagree about whether FAQ schema works?

Because everyone is partly right, which means everyone is partly wrong. "It's dead" refers to the rich result, which is true. "It's essential for AI" refers to the extractable structure, which is also true. The two claims are about different things, and the confusion comes from treating a visible search feature and machine-readable structured data as the same subject.

How should I write FAQs in 2026?

For humans and AI rather than for a rich result. Use real questions customers actually ask, and do not invent filler questions to pad the page. Put the core answer in the first line of each response, because an AI extracting the passage should get the substance immediately rather than after two sentences of throat-clearing.

Should I delete my FAQ section now that rich results are gone?

No. One professional services firm nearly deleted an FAQ section added years earlier purely to chase the rich-result trick, on the assumption that it had become worthless. The visible reward disappeared; the structural benefit did not. A clean question-and-answer page is now one of the surer routes to being quoted by AI assistants.

Does the FAQPage markup itself earn AI citations?

Not on its own, and this distinction matters. The JSON-LD does not cause ChatGPT, Perplexity, or Google's AI Mode to cite you. What earns the citation is the content shape the markup describes: a specific question with a self-contained, accurate answer. The schema helps machines parse that relationship; it does not substitute for it.

When did Google change its treatment of FAQ rich results?

In three steps. August 2023 restricted them to authoritative government and health websites, ending the tactic for ordinary businesses. March 2026 narrowed eligibility further. 7 May 2026 retired the feature entirely, for every site, with no exceptions. Anyone still adding FAQ markup expecting a SERP enhancement is working from a pre-2023 playbook.

Can an FAQ Page Get You Quoted by AI in 2026? FAQ Schema, Explained

If you have a question-and-answer section on your website, or someone once told you to add "FAQ schema" to it, you have probably heard conflicting stories about whether it still does anything for you. Ask ten people in 2026 and you will get, I promise you, at least four different answers: "it's dead," "it never worked," "it's essential for AI," "Google penalizes it now." Everyone is partly right, which means everyone is also partly wrong, and that is exactly the sort of muddle that leads business owners to either waste effort or throw away something genuinely useful.

So let us do this properly, starting with a quick translation. "Schema," or schema markup, is simply a hidden label in your page's code that tells Google and AI systems exactly what your content is; FAQ schema tells them "this is a question, and this is its answer." Now let us separate what actually happened from what people feel happened, and then answer the real question of 2026: not "does an FAQ still make my Google listing bigger?" but "does a clean question-and-answer page help an AI assistant find, trust, and quote my business?" These are not the same question, and the difference is the whole point.

TL;DR: Google removed FAQ rich results, those expandable questions beneath a search listing, in stages: back in 2023 it restricted them to authoritative government and health sites, then narrowed eligibility further in March 2026, and on May 7, 2026 it retired the feature completely, for every site, with no exceptions left. So FAQ schema no longer earns you visible real estate in the blue links, anywhere. However, the FAQPage type is still valid schema, Google still parses it to understand your page, and most importantly, a clean question-and-answer structure is one of the most extractable, most citable content shapes for AI assistants like ChatGPT, Perplexity, and Google's AI Mode. The format did not die. Its job changed.

What happened to those expandable questions under Google results?

Let us fix the timeline, because this is where most of the confusion lives. In August 2023, Google published a change to how it treats FAQ (and HowTo) rich results. Instead of showing those familiar expandable question-and-answer accordions beneath a search result for anyone who added the markup, Google restricted them, as its Search Central announcement described, to well-known, authoritative government and health websites. For everyone else, the visible feature simply stopped appearing.

This was a genuine shock at the time, because FAQ rich results had been, for a few years, a beloved little trick: a way to make your listing physically bigger in the results and push competitors down the page. As the retrospective in Search Engine Land puts it, this was the "rise and fall" of FAQ schema as an SEO tactic. Google kept narrowing the feature after that, restricting it further in March 2026, and on May 7, 2026 it retired FAQ rich results entirely. The visible FAQ enhancement does not exist anymore, for any site, government or health included.

So if your entire reason for adding FAQ markup was to grab more space in Google's search results, that reason is gone. It must be said plainly, so nobody wastes time chasing a feature that has been retired.

So is FAQ schema completely dead? No, and here is why

Here is the crucial distinction that the "FAQ is dead" crowd misses. There is a difference between a visible search feature and structured data that machines read. Google removed the feature. It did not stop reading the markup.

FAQPage remains a perfectly valid Schema.org type. Google continues to parse it to help understand what a page is about. And this is the part that matters most in 2026: the underlying shape of an FAQ, a clear question followed immediately by a direct, self-contained answer, happens to be exactly the shape that AI systems find easiest to extract and cite.

Why? Because AI assistants do not read your page the way a human skims it, top to bottom, forming a general impression. They retrieve at the passage level; they pull individual paragraphs and sections and assess each one for how well it answers a specific question. A well-formed FAQ entry is, in essence, a pre-packaged passage: here is the question a real person asks, here is a complete answer that stands on its own without needing the rest of the page for context. That is a gift to a machine trying to find a citable, self-contained response.

Does FAQ content actually get you quoted by ChatGPT and friends?

Yes, though we must be careful about why. It is not the FAQPage JSON-LD markup, on its own, that earns you a citation in ChatGPT or Perplexity or Google's AI Mode. These systems primarily read the visible text of your page. What helps is the discipline the FAQ format imposes on your writing.

When you write a real FAQ section, you are forced to do several things AI systems reward:

You phrase a heading as the exact question a human would type or speak, which matches the way people query AI assistants conversationally.
You place a direct answer immediately underneath, in the first sentence, rather than burying it three paragraphs down.
You make each answer self-contained, so it can be lifted out and quoted without the surrounding page.
You cover the specific, "long-tail" questions that broad marketing pages tend to skip.

That combination, question-shaped headings, answer-first paragraphs, and self-contained passages, is precisely what makes content extractable. The FAQ format is simply a reliable way to enforce it. And if you also add the FAQPage structured data, you give Google's parsers an extra, explicit signal about the question-answer relationship, which certainly does no harm. So: add the markup, but understand that the real work is in how you write the answers.

If schema in general still feels intimidating, our plain-language walkthrough on adding schema markup for a small business in fifteen minutes is the friendliest place to start.

How should you use FAQ schema in 2026, practically?

Here is the approach I now recommend, step by step:

Write FAQs for humans and AI, not for a rich result. Choose the real questions your customers ask, the ones they type into Google and speak into ChatGPT. Do not invent filler questions to pad the page.
Answer in the very first sentence. State the direct answer immediately, then elaborate. An AI extracting your passage should get the core answer in the first line.
Keep each answer self-contained. It should make sense pulled out on its own, with no "as mentioned above."
Add FAQPage structured data anyway. It is still valid, Google still parses it, and it clarifies the question-answer relationship for machines. Just do not expect a visible rich result from it.
Do not stuff or fake it. Marking up content as an FAQ that is not genuinely a question and answer has always been against Google's guidance and helps nobody.
Pair it with your business and local schema. For local businesses especially, FAQ content works best sitting alongside proper business markup; our local business schema guide covers how to set that foundation correctly.

A short story about a page that got quoted

A small professional-services firm I spoke with had, years ago, added an FAQ section to a service page purely to chase the rich-result trick. When Google removed the feature, the owner assumed the whole thing had become worthless and nearly deleted it. Fortunately, he did not.

Some months later, curious, he asked an AI assistant a very specific question about the service his firm offered, the kind of narrow, practical question a real prospect might ask. The assistant answered it accurately and cited his firm's page as a source. When he checked, the answer the AI had given was, almost word for word, the answer he had written in that "worthless" FAQ years earlier. The visible search feature had died. The content, written cleanly as a self-contained question and answer, had quietly become the very thing an AI chose to quote. He kept the section. Naturally.

Not sure whether your pages are structured in a way AI assistants can actually read and cite? Check your SEO Standings to see how your site looks to both Google and the new generation of AI search, or let our done-for-you SEO team handle the structure for you.