Here is the short answer: a visible FAQ section can make it easier for an AI answer engine to find and use a specific answer on your page, but it does not earn you a citation by itself, and Google's old FAQ rich snippet feature does not exist anymore. If you have been told that adding FAQ schema or a question-and-answer block is a reliable way to get pulled into ChatGPT or Google AI Overviews, that claim is going further than the evidence actually goes.
This matters because a lot of teams are still adding FAQ sections for a reason that stopped being true. Let's sort out what an FAQ section can realistically do for you, what it cannot, and how to decide whether one belongs on a given page.
Three things people mix up when they talk about FAQs
When someone asks "does an FAQ help with AI search," they are usually talking about one of three different things, and mixing them up is where most of the confusion starts.
An FAQ section is the visible questions and answers you and your readers actually see on the page. This is the thing this article is about: is that visible content worth adding.
FAQ schema is the machine-readable markup behind the scenes that describes that content to search engines. It is a separate technical layer, and this piece is not going to walk you through implementing it. That is a different question with its own answer.
An FAQ rich result, what most people mean when they say FAQ rich snippets, was Google's old expandable question-and-answer treatment that used to show up directly in regular search results. That feature is not the same thing as an AI-generated answer or an AI citation, and as you will see in the next section, it is not something you can get anymore no matter how you mark up your page.
Keep those three separate as you read the rest of this, because a lot of outdated advice about FAQ page SEO quietly swaps one for another and ends up promising something nobody can deliver. FAQ page SEO, in the broader sense of ranking a dedicated FAQ page, is also its own separate topic. This piece is narrower than that: it is only about whether a section within an existing page earns anything with AI answer engines.
Google's FAQ rich results are gone
If part of your reason for adding an FAQ section is to earn that expandable snippet in Google Search, you can stop chasing it. Google says its FAQ rich results stopped appearing in Search as of May 7, 2026. It also removed the related documentation, the rich-result report, and testing support for the feature over the months that followed, and pulled Search Console API support for it a couple of months after that.
This was not a sudden reversal. Google narrowed FAQ rich snippets back in August 2023, limiting them to well known government and health sites. A lot of guidance still floating around online reflects that earlier, narrower rule, which makes it easy to assume a properly marked up commercial page might still qualify under the old exception. It does not. The 2026 change removed the feature outright, for every kind of site, not just the ones outside that original carve out. If you still see a blog post or a forum thread promising FAQ rich snippets for an ordinary commercial page, it is describing a feature that stopped existing months ago.
So if a plan to chase AI citations is actually built on getting an FAQ rich result back, it is built on a feature that no longer exists. That distinction matters because the rest of this article is about a genuinely separate question: whether an FAQ section helps with AI answers, not classic Google search results. Losing that display feature says nothing, one way or the other, about whether AI engines find visible question-and-answer content useful when they build an answer, which is the actual question worth spending time on.
What the AI platforms actually say about FAQ sections
This is where it gets more useful, and also where you need to read the actual wording carefully instead of the version that gets repeated secondhand.
Google's own guidance for AI Overviews and AI Mode says there are no additional requirements or special optimizations needed to appear in those features, and that existing SEO best practices already apply. It also says no special schema.org markup is required. Google separately notes that these AI features may use something called query fan out, where one question gets broken into several related searches across different subtopics and sources. That helps explain why a page can be genuinely useful for a narrow question without that proving an FAQ block earned it a citation, or that FAQ format was what the system was looking for.
Microsoft is more direct, though still measured. In a February 2026 announcement about AI Performance reporting in Bing Webmaster Tools, Microsoft said plainly that FAQ sections help surface key information. Its broader webmaster guidance also emphasizes clear, accurate, and independently verifiable content that is easy to pull out of context. That is a real, specific statement in favor of the format, and it is worth taking seriously. But Microsoft did not publish a citation lift percentage or promise that an FAQ section results in more citations. It is guidance about making information easier to surface, not a measured outcome.
OpenAI's publisher guidance describes how any public website can appear in ChatGPT search and what makes content discoverable and citable, including crawler access. Nowhere in that guidance does it say an FAQ section or FAQ schema secures a citation. A 2026 cross-platform study using 602 controlled prompts across ChatGPT, Google AI Overview, Gemini, and Perplexity looked closely at how sources get selected and used in AI answers, but it did not isolate FAQ sections as a variable, so its numbers cannot honestly be quoted as an FAQ result.
Put together, the honest summary is that one platform gives a real but modest endorsement of the format, and the others describe general discoverability without singling out FAQs as a special lever. None of the three platforms above published a number you can plug into a business case, and none of them contradict each other either. They are simply answering slightly different questions: Google is talking about requirements, Microsoft is talking about surfacing, and OpenAI is talking about access. Read as a set, they support the modest version of the claim and rule out the strong one.
It also helps to know what none of this evidence covers. None of the sources above isolate an FAQ section as the single changed variable on an otherwise identical page and then measure a citation difference. That is the study that would actually settle the size of the effect, and as of this writing, nobody reviewed here has published it. The 602-prompt cross-platform study comes closest to a rigorous comparison of citation behavior, but it studies source selection broadly across ChatGPT, Google AI Overview, Gemini, and Perplexity, not FAQ formatting specifically, so its figures should not be repeated as an FAQ result, even though they get treated that way in some secondhand summaries. That gap alone is a good reason to be skeptical of any exact percentage attached to FAQs specifically.
Why an FAQ section might genuinely help, and where that reasoning stops
There is a real, sensible mechanism behind Microsoft's statement, even without a measured number attached to it. When a question and its answer sit together as a self-contained block, that block is easier for a system to identify and use in isolation, compared to the same answer buried in the middle of a longer paragraph that depends on the sentences around it. That is an explanatory inference, consistent with how these systems are described to work, not a disclosed ranking rule.
Where that reasoning stops is the leap from "easier to extract" to "more likely to be cited." Extraction, selection, and whether a link actually shows up in the answer are different steps, and a system can use information from your page without linking back to it, or cite your page for something written somewhere else on it entirely. An FAQ is one way of exposing a useful answer clearly. It is not a token that a citation algorithm rewards for existing.
That also means duplicating an answer you already state clearly in your main body, just to also restate it inside an FAQ block for citation purposes, has no demonstrated payoff. If the article already answers the question well, adding a repetitive tail section under a different heading is not a proven AI-citation tactic, it is just the same answer written twice under a different heading.
What is actually established, what is plausible, and what is still unknown
It helps to hold these at different confidence levels instead of treating them as one settled fact.
A couple of things are genuinely established. Google's FAQ rich result no longer appears in Search, as of May 2026. FAQ schema is not a special requirement for appearing in Google's AI features either, since Google's own guidance says no additional schema is needed.
One thing is plausible but not proven. A visible FAQ section can make an answer easier for an AI system to locate and use cleanly, based on Microsoft's guidance about surfacing key information. How large that effect is, or whether it holds the same way across every engine, has not been measured in a way that supports a specific number.
A few things are simply not established yet: that adding an FAQ section causes a predictable increase in AI citations, that FAQ schema is required for citation, and that the same tactic works identically across ChatGPT, Gemini, Perplexity, and Google's AI features. The engines behave differently enough that a claim proven for one should not be assumed for another.
Watch for numbers that quietly blur an FAQ section with FAQ schema, or that treat a page that happens to already be cited and also happens to have an FAQ as proof the FAQ caused it. That kind of before-and-after comparison, on its own, does not rule out the dozen other things that also changed on the page, in the index, or in the engine itself. A rich result and an AI citation are also not interchangeable measures, even though they get used that way in casual conversation. One was a display feature inside conventional Google Search. The other is a link inside an AI-generated answer, produced by a completely different system with its own selection behavior. Treating a study or a tool result about one as evidence about the other is a mistake worth catching before you build a strategy on it.
What to actually do with this
Make the call based on whether an FAQ genuinely earns its place, not based on a promised search feature that no longer exists. If a page already needs to resolve two or three distinct, recurring reader questions that do not fit cleanly into the main flow of the article, a visible FAQ can be a genuinely useful way to expose those answers. If the body of the article already answers those same questions clearly, adding an FAQ on top of that is not evidence-backed, it is just more page.
Separate the two outcomes you might be measuring. FAQ rich-result visibility in Google Search is gone, full stop, regardless of what you do. AI citation visibility is a different thing: whether an engine links to your specific page as a source when it answers a related question, not just whether it mentions your brand or echoes a sentence you wrote. Watching that second outcome over time, and connecting it back to specific pages rather than the site in general, is the only way to know whether any formatting change, FAQ included, is actually moving the number. DeepSmith's AI Visibility area tracks citations at the page level and by prompt for exactly this reason, so a team can see whether the pages they change actually start showing up as sources, instead of guessing from a single before-and-after glance.
A quick way to tell the difference: say a pricing page already covers the fee tiers clearly, but readers keep asking about cancellation in a way the page never quite addresses. Adding a short, visible answer to that specific question is a reasonable use of the format, because it resolves a real gap. Compare that to a team that tacks five generic questions onto every single post purely to reach for the old Google rich result. That treatment ended in May 2026, and the addition does nothing measurable on the AI side either. The difference is not the format, it is whether the question was actually unanswered before you added it.
If you want to build an FAQ section that reads well and is worth including for its own sake, that is a writing and structure question with its own answer elsewhere, including how to phrase the questions themselves and whether FAQ schema is worth implementing on top. This piece was only about whether the format itself earns you anything on the AI side, and the honest version of that answer is: maybe, modestly, and not automatically.
