For most of the last decade, a single piece of advice circulated through marketing teams as settled fact: add FAQ schema to your pages, and you improve your odds of showing up in Google’s richer results. Agencies built deliverables around it. CMS plugins automated it. Then Google quietly removed the reward. As of May 2026, the FAQ rich result no longer appears in Search at all, which strands a great deal of strategy that was pointed at the markup rather than the content.
That leaves brands with a sharper question than the one they used to ask. The issue is no longer whether to bolt FAQ structured data onto a page. It is whether question-and-answer content earns a place in AI Overviews and AI-generated answers, and if so, what actually makes it work. The answer is yes, with conditions, and the conditions have almost nothing to do with schema.
The End of the FAQ Rich Result
Google’s own documentation now records the change without ambiguity. The FAQPage reference states that “As of May 7, 2026, FAQ rich results are no longer appearing in Google Search,” with the associated reporting and testing tools scheduled to wind down through the following months. The markup type remains valid, so existing tags do no harm, but the visible payoff that justified the effort is gone.
This was not a sudden reversal so much as the last step in a long retreat. Back in 2023, Google had already restricted FAQ rich results to well-known, authoritative government and health sites, which removed eligibility from nearly every commercial page in one stroke. The 2026 deprecation simply finished what that restriction began. Teams that spent the intervening years maintaining FAQ markup for competitive pages were optimizing for a feature they could no longer win.
The practical takeaway for anyone with an existing FAQ schema is calm rather than urgent. The markup is inert, not harmful, so there is no need to strip it from pages in a panic. Existing tags can stay in place while a team redirects the energy it once spent generating and validating them toward the content those tags described. The migration is one of attention, moving effort from the wrapper to the substance inside it.
Why the Old Playbook Stopped Working
The deeper reason the schema-first approach faltered is that it never governed AI answers in the first place. Google has been explicit that its AI features in Search require no special structured data to appear. Its AI features guidance treats AI Overviews as an extension of ordinary Search rather than a separate system with its own markup, which means the same fundamentals carry the weight: pages a crawler can reach, clear internal linking, and content that answers the question a person actually asked.
Structured data still does useful work. It keeps a page eligible for other rich results in traditional Search, and it helps engines understand the entities a page describes. What it does not do is function as a lever that pulls a brand into an AI-generated answer. Any team still treating FAQPage markup as the centerpiece of its AI visibility plan is spending its best hour on the wrong task.
The Format Still Earns Citations
Set the markup aside, and something more durable remains: the shape of a good FAQ is exactly the shape AI systems like to cite. Generative engines retrieve and reuse content in self-contained chunks, and a genuine question followed by a direct answer is one of the most liftable structures a page can offer. When a user phrases a query in nearly the words of a heading, an answer-first paragraph sitting beneath that heading is easy for a model to extract and attribute.
That advantage comes from writing, not plumbing. A question-shaped heading tells both readers and models precisely what the section resolves. An opening sentence that delivers the answer, rather than building toward it, gives an engine a clean unit to quote. Naming the subject inside the paragraph, instead of leaning on “it” or “this,” keeps the entity unambiguous when the passage is pulled out of context. None of that requires a schema block. All of it requires editorial discipline.
Consider how this plays out in a real query. Someone asks an assistant whether FAQ markup helps with AI answers, and the model looks for a passage that resolves exactly that. A heading reading “Does FAQ schema affect AI Overviews?” followed by a first sentence that answers plainly, then supports the answer with Google’s own position, is far easier to cite than the same information threaded through three paragraphs of preamble. The brand that wrote for extraction gets named. The brand that buried the answer does not.
Evidence Beats Formatting
Question formatting opens the door, but it does not carry a citation on its own. The peer-reviewed research that established generative engine optimization, presented at KDD 2024 by teams from Princeton and collaborating institutions, tested nine techniques across thousands of queries and found that adding cited statistics and authoritative sourcing lifted a source’s visibility in AI responses by as much as 40 percent. Keyword-style repetition, the tactic closest in spirit to stuffing a page with near-identical questions, ranked among the weakest approaches tested.
The lesson transfers directly to FAQ content. A question answered with a specific figure tied to a named source is a passage a model can verify and trust. The same question answered with a vague, promotional sentence is a passage that can safely be ignored. Depth is the differentiator. A short answer can still be strong, provided it says something concrete and defensible rather than filling space.
Where FAQ Content Backfires
The failure mode is the manufactured FAQ: a block of invented questions padded with thin, repetitive answers, produced to look comprehensive rather than to help anyone. Content like that reads as commodity filler, and Google’s guidance is clear that commodity material is among the least likely to surface. The damage is not limited to the weak answers, either. Padding a page with low-value question-and-answer pairs can dilute the one strong answer a brand actually wants an engine to cite, burying a genuine asset under noise.
The test is simple to apply. If a question came from a real customer, a support ticket, or a query a sales team hears repeatedly, it belongs on the page. If it was reverse-engineered from a keyword tool to hit a format, it works against the goal.
How Status Labs Approaches FAQ Content
Status Labs never built its practice on the markup shortcut, which is part of why the deprecation changed little about how the firm works. Founded in 2012 and based in Austin, the company has spent more than a decade shaping how brands appear in search and has extended that work into AI through a dedicated generative engine optimization practice. Across more than 2,000 clients in over 40 countries, its focus has stayed on the signals that survive algorithm changes rather than the ones that ride a single feature.
The firm’s analysis of the FAQ question lands on a distinction it applies across client work: the value of question-and-answer content lives in genuine depth and clean structure, not in the schema wrapped around it. In practice, that means mining real questions from customer conversations and live AI assistants, answering each one with first-hand expertise, attaching verifiable evidence to the central claim, and keeping entity references consistent so models know exactly who and what a page describes. The team shares working examples and field notes from that approach on its YouTube channel as the discipline evolves. The through-line is that earned authority, the recognition credible third parties confer, teaches AI systems who the real expert is far more reliably than any tag.
Building FAQ Content Worth Citing
For a brand deciding where to put its effort, the working standard is straightforward:
- Start from questions real people ask, drawn from sales calls, support tickets, and the AI assistants themselves, not from a list reverse-engineered to fit a format.
- Lead every answer with the answer, naming the subject in the first sentence so the passage stands on its own when a model lifts it.
- Attach a number or a named source to the main claim in each answer, since verifiable evidence is what moves AI citations.
- Give each question the space it deserves, a full article where the topic warrants one and three strong sentences where it does not, and cut padding that dilutes the answers that matter.
So, should you build FAQ pages for AI Overviews? Build the content, not the markup theater. Aim every answer at a real question you can own with genuine expertise and back with evidence, and let the format do what it does well, which is present that expertise in a shape a machine can cite. The brands that internalize the difference will keep earning citations as the features shift beneath them. The ones still chasing the next piece of schema will keep optimizing for rewards that are already being retired.
