Authority Engine
Back to blog

FAQ Schema and SEO in 2026: What Still Works

Marketer reviewing FAQ structured data at a laptop

FAQ schema no longer earns expandable answers in Google results. As of May 7, 2026, FAQ rich results stopped appearing in Google Search, including for the government and health sites that were the last ones eligible. The markup is still valid, and the questions-and-answers content it describes is still worth writing. But the reason to do it has changed: you write FAQs for buyers and for the systems that summarize your page, not for a bigger search listing.

This guide covers what FAQ schema is, what Google changed, what the evidence says about AI search, when to keep or remove your markup, and how to write and test FAQs that hold up.

Authority Engine
Help Buyers Find You in AI Search
See how your business appears in AI-assisted buyer research, compare relevant competitors, and identify where visibility work could begin.
Get your visibility report

TL;DR

  • Google deprecated FAQ rich results on May 7, 2026, so FAQPage markup will not give you an expanded listing in standard search.
  • FAQPage is still a valid schema type. Keep it only if it is accurate, matches visible text and does not cost you engineering time.
  • A large study of pages that added JSON-LD, including FAQPage, found no meaningful lift in AI citations. Do not expect schema alone to get you cited.
  • The durable asset is the FAQ content itself: specific, honest answers to real buyer questions, visible on the page.
  • Test with Google’s Rich Results Test and URL Inspection, and never mark up text users cannot see.

Table of Contents

What FAQ schema is

FAQ schema is structured data that labels question-and-answer pairs on a page so machines can read them without guessing. Google defines structured data as a standardized format for providing information about a page and classifying its content. The most common way to write it is JSON-LD, a block of code that sits in the page and mirrors what a visitor reads.

Two parts matter:

  1. The visible FAQ. Headings or text that pose a question, followed by the answer, on the page itself.
  2. The FAQPage markup. A machine-readable copy of those same questions and answers.

The second must describe the first. Google states that structured data has to describe the content of the page it is added to, and it warns publishers against marking up information that users cannot see. Think of the markup as a label on a jar, not a second jar.

This is also the answer to a common confusion. “FAQ schema” and “FAQ content” are not the same thing. The content is the work. The schema is optional packaging.

What Google changed, and when

The decline happened in two steps.

August 2023. Google restricted FAQ rich results to well-known, authoritative government and health websites. For nearly every business, the expandable FAQ snippet disappeared then.

May 7, 2026. Google officially deprecated FAQ rich results, so they no longer appear in standard results for anyone, including those remaining government and health sites.

What this means in practice:

  • If your FAQ markup was producing an expanded snippet, it is not now.
  • Your rankings did not depend on that snippet. Google says structured data helps it understand a page and can enable rich results, and the second part is what went away.
  • Removing the markup is not required. Leaving it in place is not a violation as long as it is accurate and matches what users see.

Adoption has fallen back as well. In a September 2026 crawl, FAQPage markup appeared on 2.9% of 1,401,418 top-1M pages, and 16.3% of those pages also carried Product markup, which suggests much of what remains sits on ecommerce product and category pages, often as leftover template code.

Does FAQ schema help you show up in AI answers?

This is where most of the hopeful advice lives, so look at the evidence rather than the claims.

In an Ahrefs study of 1,885 pages that added JSON-LD schema, including FAQPage, there was no major citation uplift. AI Overview citations declined 4.6%, and the changes in AI Mode and ChatGPT were statistically indistinguishable from noise.

That does not prove schema is harmful. It shows that adding it after the fact is not a lever you can pull to get cited. The same source offers a sensible rule: FAQPage remains a valid, crawlable type, so the decision to keep it should rest on its usefulness to other consumers or automated systems, not on an expectation of Google rich results.

Why would the lift be small? AI models are built to read ordinary language. A clear question followed by a clear answer in plain HTML is already easy to interpret. What decides whether you are cited is closer to this:

  • Is the answer specific and correct?
  • Does your site sit among sources the system already trusts for this topic?
  • Can the crawler reach the page at all?
  • Is your company described consistently across the web?

We cover those factors in our guides to schema types that matter in AI search, how AI search platforms retrieve and cite sources and what works for AI Overviews.

Keep, fix or remove: a practical decision

Use this short test on each page that carries FAQPage markup.

Situation What to do
FAQ is visible, accurate, and the markup matches it exactly Keep it. It costs nothing and remains valid.
Markup includes questions or answers that are not on the page Fix it or remove it. This breaks Google’s rule against marking up invisible content.
Markup is on pages where FAQ is not the main focus, such as a product page with unrelated boilerplate questions Remove it. Misleading or irrelevant markup can put other search features at risk.
Your CMS or plugin adds it automatically and nobody maintains it Decide whether to switch it off. Stale answers with a machine-readable label are worse than none.
You are planning new FAQ pages Write the content first. Add markup only if your platform makes it trivial.

Do not spend a sprint ripping out working code. Practitioners have noted that platforms are slow to remove obsolete markup, and large sites still ship FAQPage code months after it stopped producing any visible benefit. That is a maintenance cost, not an emergency. Audit it when you are already touching those templates.

How to write FAQ content that earns its place

Since the content is the real asset, this is where the effort belongs.

Start from questions buyers actually ask. Pull them from sales calls, support tickets, onboarding emails, and the comparison questions prospects raise before they ever contact you. A question like “What does this cost and what drives the price?” beats “Why is our company the best?”

Answer in the first sentence. Give the direct answer, then the condition or nuance. A reader scanning, and a model summarizing, both need the answer before the caveat.

Keep each answer self-contained. If an answer only makes sense after reading three other paragraphs, it will not travel well when quoted. Name the subject instead of writing “it” or “this.”

Be specific and honest. Include real numbers, real limits and who the answer does not apply to. Vague answers are the ones that get skipped.

Put FAQs where the decision is made. A short block on a service page, pricing page or comparison page does more than a standalone FAQ page nobody visits.

Keep it short. Four to eight strong questions per page beat thirty thin ones. Duplicate or near-duplicate questions across many pages add noise.

Maintain it. When pricing, policy or product details change, the FAQ is the first place a stale answer hurts you.

Seven practices for useful buyer-focused FAQs

This is the same discipline behind buyer-intent content for ChatGPT: clear answers to the questions buyers ask while they are still comparing options.

How to add FAQPage markup correctly

If you decide to keep or add it, the process is short.

  1. Write the visible FAQ first. Each question and answer must be readable text on the page.
  2. Create the JSON-LD. Add FAQPage structured data as JSON-LD, with each question as a Question entity and each answer as an Answer entity. One generator guide places it in the page’s head section and stresses that the same questions and answers must also appear as visible text.
  3. Match the text exactly. If you edit an answer on the page, edit it in the markup too.
  4. Test it. See the next section.
  5. Re-check after template changes. Redesigns and plugin updates are the usual cause of drift.

A minimal example looks like this:

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [{
    "@type": "Question",
    "name": "Your question as it appears on the page?",
    "acceptedAnswer": {
      "@type": "Answer",
      "text": "Your answer, identical to the visible text."
    }
  }]
}

Wrap it in a JSON-LD script tag in your template or through your CMS’s structured data field. Do not paste in answers you have not published on the page.

How to test and monitor FAQ schema

Google’s own tools are the right check, and they cover the basics of valid syntax and crawler access.

  • Rich Results Test. Use it on live or draft pages to see how structured data parses.
  • URL Inspection in Search Console. It shows structured data errors on a specific URL and lets you verify a fix.
  • Issue details in Search Console. After you fix a problem, monitor validation progress there.

Google also advises using the Rich Results Test and URL Inspection to validate your implementation and confirm pages are accessible to crawlers. Keep in mind that a page can pass validation and still be invisible if crawlers cannot reach it, so a regular technical SEO crawl that checks schema, indexability and page structure catches problems a single-page test will not.

Because FAQ rich results are gone, a clean test result no longer means a visible reward. It means the markup is valid and not a liability.

Mistakes and policy risks to avoid

Google’s general structured data guidelines draw the lines clearly.

  • Hidden content. Do not describe content that users cannot see. Accurate but invisible markup still breaks the principle.
  • Misleading or irrelevant markup. Schema that does not fit the page’s main focus can trigger manual actions that remove eligibility for other search features.
  • Promotional filler as questions. Questions nobody asks, written to carry keywords, make the page worse for readers and add nothing for machines.
  • Copy-pasted FAQs across dozens of pages. The same block everywhere dilutes each page and increases the maintenance burden.
  • Treating schema as a citation strategy. The evidence above says it is not one.

Our perspective: schema is a label, authority is the product

We work with established companies whose buyers research and compare before they ever speak to sales. When FAQ rich results disappeared, many marketing teams asked whether the whole effort had been wasted. We think it was the wrong question.

The reason FAQ content works was never the snippet. It was that buyers have specific questions at specific moments, and the company that answers them clearly, consistently and in places AI systems trust gets considered. Schema labels that content. It does not create trust, and it does not create the external credibility that makes an AI system comfortable naming you.

Our perspective: schema is a label, authority is the product

So our advice is plain. Keep your valid markup if it is cheap to maintain. Stop expecting it to carry weight. Put your effort into three things: buyer-question research so your FAQs address real decisions, content that answers those questions better than the alternatives a buyer will see, and a consistent, verifiable description of your company wherever it appears. Our guides on organization schema and brand knowledge graphs go deeper on that last point, because conflicting facts about a company are a bigger risk in AI answers than a missing FAQ block.

GEO and AEO add to SEO. They do not replace it, and a technical tag does not substitute for either. If you want the broader view, start with our overview of generative engine optimization.

— Dr. Patrick McAvoy

Turning FAQ content into buyer consideration

If your FAQs are written but you cannot tell whether buyers see your company when they ask AI tools to compare providers, that is the gap we work on. Our Managed AI Visibility service covers buyer-question research, strategic content, external authority building and ongoing monitoring, so the work does not stop at a markup fix. To see where you stand first, request the AI Visibility & Revenue Opportunity Report.

Authority Engine

What to do this month

  1. List every template that outputs FAQPage markup, including plugin-generated ones.
  2. Compare the markup to the visible text on a sample of pages. Fix or remove mismatches.
  3. Remove FAQ markup from pages where FAQs are boilerplate and not the main focus.
  4. Rewrite your five highest-traffic or highest-intent FAQ blocks using real buyer questions and answer-first sentences.
  5. Run the Rich Results Test and URL Inspection on the pages you changed.
  6. Review whether your company is described the same way on your site, profiles and third-party pages.

None of this needs a rebuild, and all of it holds up whatever Google changes next.

FAQ

Does FAQ schema still work for SEO?

Not for rich results. Google stopped showing FAQ rich results in standard search on May 7, 2026. The markup is still valid and can still describe your content to systems that read it, but it is not a ranking lever and should not be sold as one.

What is FAQ schema?

FAQ schema, or FAQPage structured data, is code that labels question-and-answer pairs on a page. It is usually written as JSON-LD. The questions and answers it lists must also appear as visible text, because Google says structured data should describe what users can see.

Are FAQs good for SEO?

FAQ content is good for readers, and that is what carries over. Clear answers to real buyer questions help people decide and give search and AI systems quotable, self-contained text. Thin, repetitive or keyword-stuffed FAQs help no one, with or without markup.

Does FAQ schema help with AI search and citations?

The available evidence says not much. A study of 1,885 pages that added JSON-LD, including FAQPage, found no major citation uplift. Citations depend more on answer quality, source trust and whether crawlers can reach your pages.

Should I remove FAQ schema from my site?

Only if it is inaccurate, hidden from users, irrelevant to the page or costly to maintain. Accurate markup that matches visible text is harmless to keep. Review it when you next update the templates that produce it.

Sources