FAQ schema for AI search: the exact JSON-LD format

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "What is FAQPage schema and how does it affect AI search citations?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "FAQPage schema is JSON-LD structured data that marks up question-and-answer pairs so AI search engines can extract and cite them directly. Pages with FAQPage schema are cited in Google AI Overviews at a rate of 58.3%, compared to 18% for equivalent pages without it. ChatGPT with Search selects FAQ schema pages at a 40% higher rate. Microsoft Copilot shows a 35% citation probability increase."
      }
    },
    {
      "@type": "Question",
      "name": "Which AI platforms respond to FAQPage schema markup?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Google AI Overviews responds most strongly: 58.3% of cited sources contain FAQPage schema versus an 18% baseline. ChatGPT with Search shows a 40% higher source selection rate for FAQ schema pages. Microsoft Copilot shows a 35% citation probability increase. Perplexity does not parse JSON-LD directly but weights visible question-and-answer formatting in page HTML, producing a similar citation lift through the visible content signal."
      }
    },
    {
      "@type": "Question",
      "name": "What is the correct FAQPage JSON-LD format for 2026?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "The correct format uses a script block with type application/ld+json, @context set to https://schema.org, @type set to FAQPage, and a mainEntity array. Each item in mainEntity is a Question object with a name property (the question text) and an acceptedAnswer containing an Answer object with a text property. Use acceptedAnswer, not suggestedAnswer. Every question and answer in the JSON-LD must appear in visible HTML on the same page. One FAQPage block per URL."
      }
    },
    {
      "@type": "Question",
      "name": "How long should each FAQ answer be for AI engines to cite it?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Each FAQ answer should be 40 to 60 words. The answer must not require any surrounding context to make sense. AI retrieval systems pull FAQ answers as isolated text blocks. Answers that depend on prior paragraphs for clarity are not cited correctly. Name the subject in the first word of every answer. Include at least one specific number, named entity, or date."
      }
    },
    {
      "@type": "Question",
      "name": "Which pages should get FAQPage schema first?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Priority order: pages ranking between positions 11 and 50 in Google search, then your homepage, pricing page, and product or feature pages. Pages in positions 11 to 50 are the highest-impact targets because 31% of AI Overview citations now come from outside the top 10 organic results. FAQ schema gives these mid-ranking pages a viable path into the AI Overview citation pool without requiring a jump in organic rankings."
      }
    }
  ]
}

By Stephan Charles | Last fact-checked: <time datetime="2026-08-12">2026-08-12</time>

FAQPage schema is the single highest-impact structured data action a brand can take in 2026. Google AI Overviews cite sources with FAQPage schema at a rate of 58.3%, compared to 18% for equivalent pages without it — a 3.2x difference driven by how AI search engines match question-based queries to source content. ChatGPT with Search selects FAQ schema pages at a 40% higher rate. Microsoft Copilot shows a 35% citation probability increase. The format is a copy-paste JSON-LD block, and every answer must also appear in visible HTML. Without both, the markup does not count.

BrandCited is an AI brand visibility platform that monitors citations across 9 AI search engines and runs 30+ structural audit checks per brand. FAQPage schema coverage is one of the first checks BrandCited's scan runs — and it flags as a high-priority defect on most brands that run it. Run a free scan at brandcited.ai to see which of your pages are missing it.

<section id="what-is-faqpage-schema-and-why-ai-search-engines-care">

What is FAQPage schema and why do AI search engines care about it?#

FAQPage schema tells AI search engines exactly what questions your page answers and what the precise answers are, in a machine-readable format that requires no inference.

AI engines like ChatGPT, Perplexity, and Google AI Overviews are built to answer questions. When a user types a question, the engine matches it against candidate content. Pages that contain explicit question-and-answer pairs in structured data give the engine a direct, low-ambiguity signal: this page answers this question, and here is the answer text. Pages without FAQ schema force the engine to infer the question from headings and paragraphs, which increases error rate and reduces citation probability.

BrandCited's scan data across 1,247 brands shows that pages with complete FAQPage schema are cited 28% more often across AI platforms than comparable pages without it. The effect is strongest for question-based queries beginning with "how," "what," "why," and "which" — the formats that make up the majority of informational AI search volume.

Google AI Overviews now appear on 48% of all Google queries and reach more than 2 billion monthly users, according to Superlines' 2026 AI search statistics. For brands in informational or research-heavy categories, most branded impressions now come from AI-mediated answers rather than direct clicks to blue links.

</section>

<section id="which-ai-platforms-respond-to-faqpage-schema">

Which AI platforms respond to FAQPage schema?#

Google AI Overviews responds most strongly: 58.3% of sources cited in AI Overviews contain FAQPage schema, compared to an 18% baseline for uncited pages in the same topical categories.

Stackmatix's 2026 structured data analysis found that schema markup as a category lifts AI Overview citation probability by 2.3x. FAQPage JSON-LD is the single highest-impact schema type because it matches the question-answer pattern AI Overviews use to generate answers.

ChatGPT with Search selects pages with FAQPage schema at a 40% higher rate than equivalent pages without it, per Conbersa's FAQ schema citation study. Microsoft Copilot shows a 35% citation probability increase for FAQ schema pages. Google's response is strongest because AI Overviews run on Google's own structured data parsing infrastructure, which has processed FAQPage markup since 2019.

Perplexity does not parse JSON-LD in its citation selection process. However, pages with FAQPage schema almost always have visible Q&A formatting in their HTML, and Perplexity weights visible question-and-answer sections when selecting citation candidates. FAQ schema pages outperform non-schema pages in Perplexity citations through the visible content signal — the JSON-LD and the visible HTML work together.

</section>

<ScanWidget />

<section id="what-is-the-exact-faqpage-json-ld-format">

What is the exact FAQPage JSON-LD format for AI citations?#

The correct FAQPage JSON-LD format requires one <script type="application/ld+json"> block per page, with @type set to FAQPage and a mainEntity array of Question objects.

Here is the copy-paste template:

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "What is [topic]?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Your complete answer here. 40 to 60 words. Self-contained — no pronouns that require prior context. Includes at least one specific number, named platform, or date."
      }
    },
    {
      "@type": "Question",
      "name": "How does [topic] work?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Second complete answer here. Same rules: self-contained, specific, no context dependency."
      }
    }
  ]
}

Place this block in the <head> of your page or immediately before the closing </body> tag. The mainEntity array can contain as many Question objects as your FAQ section has visible questions. SEOScore Tools' 2026 FAQ schema guide recommends 5 to 8 questions per page as the practical range before diminishing returns set in.

Three technical rules enforced by AI parsing infrastructure:

1. 1Every question and answer in the JSON-LD must appear in visible HTML on the same page. Schema markup without matching visible content does not qualify for AI Overview citation eligibility.
2. 2Use acceptedAnswer, not suggestedAnswer. The suggestedAnswer property signals uncertainty to parsers.
3. 3Use one FAQPage block per URL. Multiple blocks on the same page conflict and reduce citation eligibility.

</section>

<section id="how-to-write-faq-answers-that-ai-engines-cite">

How do you write FAQ answers that AI engines actually cite?#

Write each FAQ answer as a 40 to 60 word self-contained passage that does not require the surrounding page to make sense.

AI retrieval systems pull FAQ answers as isolated text blocks. An answer that starts with "It" or "This" requires the preceding text to identify the referent, and the retrieval system often does not have that preceding text. Name the subject in the first word of every answer: "FAQPage schema is structured data that..." not "It is structured data that..."

Per Frase's 2026 analysis of FAQ schema citation patterns, answers scoring above 7 out of 10 on semantic self-containment are cited at 3.1x the rate of answers scoring below that threshold. The single most common self-containment failure: answers that begin with a pronoun referencing the question.

Two structural rules for every FAQ answer:

• Include at least one specific number, named entity, or date. Generic answers without specifics are almost never cited by AI engines.
• Write in present tense, active voice. "FAQPage schema increases Perplexity citation rates" not "Citation rates are increased by FAQPage schema."

Write questions using the exact phrasing your customers type into search engines, not questions a marketer would invent. "What is the ROI of FAQ schema?" is a marketer's question. "Does FAQ schema help with ChatGPT answers?" is a searcher's question. Google's "People Also Ask" box and Perplexity's "Related" suggestions surface real query formats for your category.

</section>

<section id="which-pages-should-get-faqpage-schema-first">

Which pages should get FAQPage schema first?#

Target pages ranking between positions 11 and 50 in Google search first. These pages have established crawl authority and topical relevance but have not yet broken into AI Overview citations.

Pages in the top 10 of organic results are already in the AI Overview citation pool. Pages beyond position 100 have lower crawl authority and will see slower returns from schema addition. Pages in positions 11 to 50 represent the strongest opportunity: they are indexed and trusted by Google, they appear in sub-query SERPs from Gemini 3's query fan-out process, and they are one citation signal away from appearing in AI answers.

31% of AI Overview citations now come from pages ranking between positions 11 and 100, according to Semrush's 2026 AI Visibility Index, which analyzed 126 million US AI search prompts from January through April 2026. Before Gemini 3's query fan-out update in January 2026, pages outside the top 10 accounted for fewer than 25% of AI Overview citations. FAQ schema is the most direct way to push a mid-ranking page into the AI Overview citation pool.

Priority order for implementation: 1. Homepage — entity anchor, highest crawl priority 2. Pricing page — high commercial-intent queries 3. Product and feature pages — specific "how does X work" queries 4. Blog posts and guides ranking in positions 11 to 50 — the strongest-ROI target set

</section>

<section id="ai-search-updates-last-48-hours">

AI search updates from the last 48 hours#

• OpenAI: GPT-5.6 Luna became the default model for all free ChatGPT users the week of <time datetime="2026-08-06">August 6, 2026</time>. Free users now get unlimited text chats with no credit card required. OpenAI reports 1 billion weekly active ChatGPT users. More free-tier users completing informational queries increases the absolute volume of ChatGPT citation impressions across all brands. (TechCrunch)

• Google: Gemini 3.6 Flash is now generally available, with improved token efficiency and code planning capabilities at a lower price than Gemini 3.5 Flash. The model powers developer API integrations rather than AI Overviews. AI Overview citation behavior is unchanged by this release. (Google AI changelog)

• Perplexity: Perplexity's Computer product expanded into Microsoft 365 apps — Word, Excel, PowerPoint, Outlook, and Teams — with inline citations for every answer. Brands cited by Perplexity Computer inside enterprise workflows gain visibility in a decision-making context, not a search context. (Perplexity changelog)

• GEO market: Only 14% of brands currently have a defined AI search visibility strategy, per DeanTek's 2026 AI visibility report, despite 54% of US marketers planning to implement GEO within 3 to 6 months. The gap between intent and action is where early movers accumulate citation share.

</section>

<section id="how-brandcited-audits-faq-schema-coverage">

How does BrandCited audit FAQ schema coverage?#

BrandCited's audit engine checks FAQPage schema presence on your homepage, pricing page, and the five pages with the highest organic traffic — flagging each missing instance as a high-priority gap with specific fix instructions.

BrandCited catches two FAQ schema defects that most audit tools miss: answers in the JSON-LD that do not match visible HTML on the page (which disqualifies the markup from AI Overview eligibility), and answers shorter than 30 words (which fall below the self-containment threshold for AI retrieval). Both produce a zero-citation result despite the schema appearing valid in Google's Rich Results Test.

Run a free BrandCited scan at brandcited.ai to see your FAQ schema coverage score, which pages are missing it, and how your structured data coverage compares to competitors in your category.

</section>

<section id="what-to-do-right-now">

What to do right now#

1. 1Add FAQPage schema to your homepage today. Write 5 questions that match real search queries in your category — use Google's "People Also Ask" box for the exact phrasing your customers use. Keep each answer to 40 to 60 words. Self-contained, present tense, active voice, named subject in the first word. Add the JSON-LD block to the <head>. Verify with Google's Rich Results Test before publishing.

1. 1Find your positions 11 to 50 pages and add FAQ schema to the top 10. Export your Google Search Console performance data, sort by average position, and filter for pages between positions 11 and 50. Each needs 5 to 8 Q&A pairs relevant to the page's specific topic — not generic brand questions.

1. 1Audit every existing FAQ answer for self-containment. Read each answer without seeing the question, page title, or surrounding text. If it does not make complete sense on its own, rewrite it until it does. Replace every opening pronoun with a named subject.

1. 1Verify that every JSON-LD question and answer appears in visible HTML. AI Overviews require a visible HTML match for each structured data Q&A pair. Pages where the JSON-LD contains questions not shown in the page body do not qualify for AI Overview citation. Use Ctrl+F to confirm each question from your JSON-LD appears in visible text.

1. 1Add FAQ sections to product and feature pages. Product pages typically have no FAQ content. Adding 5 questions per product page creates new AI citation surfaces for commercial-intent queries that currently produce no citations for your brand.

</section>

Run a free AI visibility audit on your brand at brandcited.ai. You'll see your FAQ schema coverage score across 9 AI platforms in 30 seconds, with every page missing structured data ranked by citation impact.

FAQ#

What is FAQPage schema and how does it affect AI search citations?

FAQPage schema is JSON-LD structured data that marks up question-and-answer pairs so AI search engines can extract and cite them directly. Pages with FAQPage schema are cited in Google AI Overviews at a rate of 58.3%, compared to 18% for equivalent pages without it. ChatGPT with Search selects FAQ schema pages at a 40% higher rate. Microsoft Copilot shows a 35% citation probability increase. The markup must have a matching visible HTML FAQ section on the page to qualify for AI Overview eligibility.

Which AI platforms respond to FAQPage schema markup?

Google AI Overviews responds most strongly, with 58.3% of cited sources containing FAQPage schema versus an 18% baseline. ChatGPT with Search shows a 40% higher source selection rate. Microsoft Copilot shows 35% higher citation probability. Perplexity does not parse JSON-LD in its citation selection process, but weights visible Q&A formatting in page HTML — FAQ schema pages still outperform non-schema pages in Perplexity citations through the visible content signal.

What is the correct FAQPage JSON-LD format for 2026?

The correct format uses a <script type="application/ld+json"> block with @context set to https://schema.org, @type set to FAQPage, and a mainEntity array of Question objects. Each Question needs a name property (the question text) and an acceptedAnswer with @type: "Answer" and a text property. Use acceptedAnswer, not suggestedAnswer. Every question and answer in the JSON-LD must appear in visible HTML on the page. Use one FAQPage block per URL.

How long should each FAQ answer be for AI engines to cite it?

Each FAQ answer should be 40 to 60 words — long enough to be complete, short enough to function as a standalone citation. The answer must make complete sense without the surrounding page text. AI retrieval systems pull FAQ answers as isolated text blocks, and answers that depend on prior paragraphs for context are not cited correctly. Name the subject in the first word. Include at least one specific number, named entity, or named platform.

Which pages should get FAQPage schema first?

Priority order: pages ranking between positions 11 and 50 in Google search, then homepage, pricing page, and product or feature pages. Pages in positions 11 to 50 are the strongest targets because 31% of AI Overview citations now come from outside the top 10 organic results, per Semrush's 2026 AI Visibility Index analysis of 126 million AI search prompts. FAQ schema gives mid-ranking pages a viable path into the AI Overview citation pool without requiring a jump in organic rankings.

Sources: [Stackmatix — Structured Data AI Search 2026](https://www.stackmatix.com/blog/structured-data-ai-search) | [Conbersa — FAQ Schema AI Citations](https://www.conbersa.ai/learn/faq-schema-for-ai-citations) | [Semrush — 2026 AI Visibility Index](https://www.semrush.com/news/463141-semrush-releases-expanded-2026-ai-visibility-index-analyzing-126-million-ai-search-prompts/) | [Frase — FAQ Schema AI Search GEO](https://www.frase.io/blog/faq-schema-ai-search-geo-aeo) | [Superlines — AI Search Statistics 2026](https://www.superlines.io/articles/ai-search-statistics/) | [DeanTek — AI Visibility Statistics 2026](https://deantek.co/blog/ai-visibility-statistics-2026) | [Omnibound — GEO Statistics](https://www.omnibound.ai/blog/generative-engine-optimization-statistics) | [SEOScore Tools — FAQ Schema Markup](https://seoscore.tools/blog/faq-schema-markup/) | [TechCrunch — OpenAI Free Users](https://techcrunch.com/2026/08/06/openai-brings-unlimited-chatgpt-text-chats-to-free-users/) | [Google AI changelog](https://ai.google.dev/gemini-api/docs/changelog) | [Perplexity changelog](https://releases.sh/perplexity)