Google Killed FAQ Rich Results. Don’t Let That Kill Your FAQ Page.

As of May 7, 2026, Google no longer supports FAQ rich results. We all got very used (and with a good reason) to those expandable question-and-answer dropdowns that used to appear beneath certain search listings, giving pages a larger SERP footprint and, for a while at least, a measurable bump in click-through rate. Search Console will stop reporting on FAQ structured data entirely, support in the Rich Results Test disappears in June, and API support follows in August. The announcement was brief, and the SEO community responded the way it always does when a familiar feature gets pulled: with a wave of tactical questions about whether to delete the markup, whether rankings would be affected, and whether FAQ pages as a content format still had any purpose at all.

These are the wrong questions (in my humble opinion), since they are tactically focused on a feature that was, in the grand scheme of things, a minor convenience, and they distract from something far more consequential that FAQ pages have been doing, largely unnoticed, for anyone who built them with genuine intent rather than schema bait.

How we ended up here, and why it matters that we understand it

To understand what is actually at stake with this announcement, it helps to understand how FAQ rich results became so ubiquitous in the first place, because the story of their rise is also the story of why so much FAQ content is so bad, and why that badness is precisely what Google wants to walk away from.

When Google introduced FAQ rich results, it created an immediate incentive that had almost nothing to do with helpfulness. SEOs noticed that pages marked up with FAQPage schema could unlock expanded listings in search, sometimes taking up significantly more visual real estate (share of voice) on the results page, and the click-through rate implications were hard to ignore. With this promise of the Holy Grail, FAQ sections proliferated across the web at a speed that had very little to do with whether those pages were answering questions anyone was actually asking and very much to do with fitting keyword targets. Answers were written to satisfy a parser, not a person. The FAQ format, which in its honest form is one of the clearest ways a brand can communicate what it does and who it serves, became a mechanical exercise in structured data configuration dressed up as content strategy.

Google removing support for this feature is, in that context, a reasonable correction. But the risk now is that the industry interprets the removal as a verdict on FAQ pages themselves, rather than on the cynical way they were so often deployed. Those are very different things, and conflating them would be a significant mistake at exactly the moment when the underlying value of a well-constructed FAQ page has become more relevant than ever.

From parsers to comprehension

The reason FAQ pages matter in 2026 has nothing to do with schema, and understanding why requires stepping back from the traditional SEO mental model (in which success means satisfying a crawler’s technical requirements and ranking signals) and into a different frame entirely, one defined by how large language models actually process and represent information about the world.

When ChatGPT, Perplexity, Google’s AI Overviews, or any other LLM-powered system tries to understand your brand, it is not checking whether you have implemented FAQPage markup correctly. It is reading your content the way a thoughtful and well-informed human reader would, absorbing language, drawing inferences, building a model of what your entity is and what it does, and then deciding, based on the clarity and consistency and specificity of what it has found, whether it can represent your brand accurately and confidently when a user asks a relevant question. Schema is a label you attach to content so that a parser can categorize it but LLMs are not looking for labels. They are looking for the kind of comprehension signals that allow them to understand, and therefore to trust, what they are reading.

This distinction matters enormously because it means that the content formats which served SEO’s parser-optimization era and the content formats which serve LLM visibility are not the same thing. And in the case of FAQ pages, what looked like a structural trick for gaming rich results turns out, when built properly, to be one of the most naturally LLM-readable formats a website can produce.

What a FAQ page actually does for entity definition and LLM understanding

Entity definition (the process of establishing a clear, consistent, machine-readable understanding of who you are, what you do, and how you relate to other entities in your domain) is the foundation of being recommended rather than merely indexed. It is the difference between an AI system that can describe your brand accurately when someone asks about your category and one that either ignores you entirely or, worse, gets you wrong in ways that erode trust without you noticing it happened. FAQ pages, when constructed with genuine intent rather than schema optimization in mind, contribute to entity definition in several distinct and meaningful ways.

Definitional clarity: saying what most websites are too vague to say

The most direct contribution is definitional clarity. Most websites bury their actual identity inside hero copy that is written to sound impressive rather than to communicate precisely through the use of language that is emotive and brand-inflected, and often deeply ambiguous about what the company actually does, who it actually serves, and what makes it different from the alternatives. An FAQ page that directly answers “What is [Brand]?”, “What does [Brand] do?”, “Who is [Brand] for?”, and “How is [Brand] different from [obvious competitor or category alternative]?” is doing foundational entity definition work that most About pages never come close to accomplishing. And LLMs need that specificity since they cannot recommend what they cannot clearly describe, and they cannot clearly describe what has never been clearly stated.

Mapping your expertise onto the question space AI is already navigating

Beyond definitional clarity, FAQ pages contribute something subtler but equally important: they map your expertise onto the question space that AI systems are navigating when they respond to user queries. LLMs generate answers by drawing on the full context of what they know about a domain: the questions that arise, the objections that are raised, the comparisons that get made, and the nuances that separate a naive understanding from an informed one. A brand whose FAQ page covers that same question surface in depth (not through keyword-stuffed approximations of real questions, but through the actual queries that appear in sales calls and customer conversations and support tickets) demonstrates a command of its own domain that LLMs can recognize and reward. It is telling the machine: we understand this space well enough to anticipate what you will be asked about it.

The structural argument: why FAQ answers are naturally extractable

There is also a structural argument, and we call it extractibility. LLMs favor content that can be surfaced confidently, which in practice means content that makes clear, complete, self-contained claims: sentences and paragraphs that hold their meaning even when extracted from their surrounding context. The FAQ format, a question followed by a direct and complete answer, is one of the few content structures that naturally produces this kind of extractable, self-sufficient language. Every well-written FAQ answer is, in effect, a passage an LLM can use without needing to import the context around it to make sense of it, which is structurally aligned with how language models actually retrieve and deploy information.

The difference between building for a rich result and building for entity comprehension

The FAQ pages that proliferated during the rich result era and the FAQ pages that actually serve entity definition and LLM visibility look quite different from each other, and the distinction is worth being explicit about because the surface format is the same even when the underlying purpose is completely different.

Rich result FAQs were optimized for a parser. They needed the right markup, the right character counts, the right structural formatting and beyond that, almost nothing else was required for them to do their job. The questions were often chosen based on keyword research rather than genuine customer inquiry, which meant they were plausible-sounding but subtly artificial, disconnected from the actual language that real people use when they are trying to understand a product or service. The answers were written to be short enough to display cleanly in a dropdown, which created a strong incentive toward incompleteness, answers that technically responded to the question without actually resolving the underlying uncertainty that prompted it. The result was a format that looked like helpfulness from the outside but was, in practice, a fairly thin layer of structured content built on a keyword and markup foundation.

An FAQ page built for entity definition starts from a completely different place. It begins with the actual questions that come up in real conversations with customers who are close to a decision but not quite there, the objections that recur across sales calls, the comparisons that prospects always seem to land on, and the definitional uncertainties that cause people to hesitate. It answers those questions with the specificity and completeness that a thoughtful expert would bring to a real conversation, not with the economy of language that a SERP feature constraint demands. It uses the brand’s own terminology consistently, so that each answer reinforces the same entity signals rather than shifting its language opportunistically to capture different keyword variations. And it is written in a way that serves a human reader first, because content that is genuinely useful to humans is, consistently and almost without exception, the content that AI systems learn to trust.

You can remove the schema. Keep the page.

Google’s own guidance on this is measured: you can remove FAQ structured data if you want, but you do not have to, because other search engines may continue to process it for their own purposes. That is sensible advice as far as it goes, but it somewhat misses the more important point, which is that the decision about whether to keep the markup is genuinely minor compared to the decision about whether to keep, improve, or abandon the underlying content.

The right question for every FAQ page on your site is not whether it has valid schema but whether it is doing real entity definition work: whether it is answering questions that your actual customers ask, in language that is clear and specific and consistent enough to give a machine everything it needs to understand and represent your brand accurately. If it is doing that work, then it is one of your most valuable GEO assets regardless of what any SERP feature does or does not do with it. If it is not (if it is a list of keyword-adjacent non-questions answered in the kind of vague, hedged, corporate-passive language that sounds like content but communicates almost nothing) then yes, it should probably be reconsidered, not because Google dropped the rich result but because bad entity signals are worse than no entity signals when you are trying to build a coherent, trustworthy brand presence that AI systems can confidently represent.

Recognition before recommendation

There is a principle that I keep returning to in almost every conversation about GEO and LLM visibility, and it applies here directly: an AI system cannot recommend a brand it does not understand. Recommendation is downstream of recognition, and recognition is built through the accumulated weight of consistent, clear, entity-defining content across your entire web presence: your website, your external mentions, your structured data where it applies, and critically, the pages on your own site that do the most direct work of explaining who you are and what you do.

FAQ pages, when built with that purpose in mind rather than with a rich result in the rearview mirror, are among the most direct and structurally natural ways to accelerate that recognition. They answer the exact kinds of questions that machines are trying to resolve when they encounter your brand. They provide the definitional specificity that allows an LLM to move from a vague category-level understanding of your domain to an accurate entity-level understanding of your brand’s particular place within it. They produce content that is, by its very structure, extractable and deployable — the kind of language that shows up in AI-generated answers because it was written to be clear enough to stand alone.

Google removing FAQ rich results changes one small part of your SERP strategy but it doesn’t change this. The FAQ page was always (first and foremost) for the understanding, and in a search environment increasingly mediated by AI systems whose entire job is to decide which brands they understand well enough to recommend, that understanding is more strategically valuable now than it has ever been. Do not let a deprecated SERP feature take it with it.