# FAQPage vs HowTo Schema: Which One Drives More AI Citations in 2026? **URL:** https://organikpi.com/blog/technical-seo/ai-search-faq-vs-howto-schema-decision/ **Published:** 2026-05-04 **Modified:** 2026-06-12 **Author:** Daniel Shashko > FAQPage and HowTo schema follow separate deprecation and AI citation tracks. Google limited FAQ rich results to government and health sites in August 2023, then removed them entirely on May 7, 2026. HowTo mobile rich results were deprecated in the same 2023 wave. Despite both removals, FAQPage schema retains full AI citation value because AI retrieval systems parse machine-readable Q&A pairs independently of SERP display rules. Our March 2026 study of 42,971 citations across 520 queries confirmed structured Q&A pages earn more citations per query. Our May 2026 study of 153,425 citations showed cited sentences average 9.27 words, matching the natural output of 40-to-100-word FAQ answers. Choose FAQPage for question-shaped content, HowTo only for genuine ordered-step procedures, and never run both primary schemas on one page. --- > FAQPage and HowTo schema follow separate deprecation and AI citation tracks. Google limited FAQ rich results to government and health sites in August 2023, then removed them entirely on May 7, 2026. HowTo mobile rich results were deprecated in the same 2023 wave. Despite both removals, FAQPage schema retains full AI citation value because AI retrieval systems parse machine-readable Q&A pairs independently of SERP display rules. Our March 2026 study of 42,971 citations across 520 queries confirmed structured Q&A pages earn more citations per query. Our May 2026 study of 153,425 citations showed cited sentences average 9.27 words, matching the natural output of 40-to-100-word FAQ answers. Choose FAQPage for question-shaped content, HowTo only for genuine ordered-step procedures, and never run both primary schemas on one page. FAQPage schema and HowTo schema serve different content shapes, and choosing wrong costs you AI citations. FAQPage is the right default for most pages because it maps each Q&A pair to the atomic chunking pattern AI engines use natively. HowTo is correct only when your content is a genuine procedural sequence with ordered, executable steps. Google deprecated both schemas from SERP rich results in August 2023, then completed the FAQ removal in May 2026, yet both schemas retained and gained AI citation value at every step of that rollback. This post explains the deprecation history, the decision framework, and the correct JSON-LD implementation for each type. ## The deprecation timeline and what it means In August 2023, Google announced that [FAQ rich results would be limited to authoritative government and health sites, and HowTo rich results would only show on desktop](https://developers.google.com/search/blog/2023/08/howto-faq-changes). The stated reason: SERP clutter and widespread abuse of the formats. Most SEOs treated this as a signal to remove the schema entirely. That was the wrong call. In May 2026, Google [completed the FAQ rich result removal entirely, dropping the search appearance for all sites](https://developers.google.com/search/docs/appearance/structured-data/faqpage). Search Console enhancement tracking and rich result testing support were also removed. The schema now has zero traditional SERP value. Yet pages with FAQPage schema continue to earn more AI citations than equivalent pages without it in our client work. The reason is structural. FAQPage schema was removed from SERP display because Google decided that real estate was better used elsewhere. The machine-readable Q&A pairs the schema creates remain exactly what [AI retrieval pipelines need](https://organikpi.com/blog/technical-seo/content-chunking-rag-seo/): discrete, self-contained chunks that can be matched to query intent and returned as cited answers. SERP display rules and AI extraction rules are separate mechanisms. ## Why FAQ schema still drives AI citations AI Overviews, Perplexity, and ChatGPT use [structured data as a retrieval signal](https://organikpi.com/blog/technical-seo/schema-markup-ai-search/), not just a SERP formatting instruction. A page with FAQPage schema provides two things to an AI retrieval system: structured query-answer pairs that require no DOM parsing to extract, and an implicit quality signal that the page has been deliberately organized to answer specific questions. Each Question entity in a FAQPage block becomes an independently citable unit. A page with 7 FAQ items is, from the retrieval system’s perspective, a page with 7 discrete answer candidates. Our [March 2026 study of 42,971 citations across 520 queries](https://organikpi.com/blog/geo-ai-search/decoded-42971-ai-citations-google-research/) covered 6 AI platforms including AI Mode, Gemini, ChatGPT, Perplexity, Copilot, and Grok. Pages with structured Q&A markup in our test set were cited significantly more often per query than comparable unstructured pages. Our [May 2026 study of 153,425 citations](https://organikpi.com/blog/seo-strategy/ai-mode-text-fragments-dead-153425-citations/) confirmed the sentence-length pattern that FAQ answers hit naturally. Cited sentences had a mean of 9.27 words and a median of 10 words. FAQ answers written at 40 to 100 words produce multiple citation-length sentences per item. That is the structural sweet spot. ## The FAQPage vs HowTo decision framework The core question: does your content describe a set of questions and answers, or a procedure with ordered, irreversible steps? Use this decision flow: Schema type decision flow for AI citations The hardest cases are content that looks procedural but is not. “Best practices for X” and “Tips for Y” lists are not HowTo content. The items are not ordered, not irreversible, and skipping item 3 does not prevent item 4 from working. Reserve HowTo for content where following step 2 before step 1 would cause the task to fail. Content typeCorrect schemaReasonGuide with Q&A pairsFAQPageMaps to query-answer chunkingTutorial with ordered stepsHowToSteps are executable in sequenceTips list or best practicesFAQPage or BlogPostingItems not ordered or irreversibleReference doc or glossaryFAQPageTerm-definition pairs match Q&A shapeRecipe or assembly instructionsHowToClear sequence, defined end state ## When to use neither Use BlogPosting or Article schema for opinion pieces, news commentary, and essays with no Q&A structure and no procedural sequence. Adding FAQPage schema to a post with no visible FAQ section violates Google’s guidelines requiring schema to match visible content, and can trigger manual actions for hidden schema. The schema must reflect what users actually see on the page. ## Implementing FAQPage: required properties The [FAQPage type on Schema.org](https://schema.org/FAQPage) is a subtype of WebPage. The minimum viable implementation requires @context, @type, and a mainEntity array of Question entities. Each Question requires a name (the question text) and an acceptedAnswer containing an Answer entity with a text property. { "@context": "https://schema.org", "@type": "FAQPage", "mainEntity": [ { "@type": "Question", "name": "Does FAQPage schema still have AI citation value after the May 2026 deprecation?", "acceptedAnswer": { "@type": "Answer", "text": "Yes. Google deprecated FAQ rich results from SERP display in May 2026, but FAQPage schema retains full AI citation value. AI retrieval systems parse the machine-readable Q&A pairs independently of SERP display rules." } } ] } Common mistakes that suppress AI extraction: putting only one Question (AI engines prefer multiple discrete units for broader query coverage), putting answer text in the Question name property instead of acceptedAnswer.text, and including questions in the schema that have no visible counterpart on the page. Schema content must match visible content exactly. We run [FAQ schema on every qualifying post](https://organikpi.com/blog/technical-seo/faq-howto-article-schema-ai-citations/) as part of our house practice, with implementation in the ACF faq_items field rendered as JSON-LD on the front end. ## Implementing HowTo: required properties The [HowTo type on Schema.org](https://schema.org/HowTo) extends CreativeWork and requires a name (the procedure title) and a step array of HowToStep entities. Each HowToStep requires a name (step title) and a text (step description). Optional but high-value: totalTime in ISO 8601 duration format, image per step, and supply or tool arrays. HowTo rich results for mobile were deprecated in the August 2023 wave. Like FAQPage, the schema is still actively parsed for [AI Overviews](https://organikpi.com/blog/geo-ai-search/google-ai-overviews-optimization-playbook/) and Perplexity procedural answers. AI engines consistently surface HowTo-marked content for procedural intent queries over equivalent unmarked content with the same information. The effect is strongest on content with 5 or more clearly sequential steps. ## Why not both on the same page Having two competing primary schema types on one page creates entity ambiguity. AI retrieval systems need to classify the page as a type; competing primary schemas force the system to guess. In practice, the page often gets parsed as neither and loses structured data coverage. If a page has both a procedural section and a Q&A section, pick the dominant type and render the other as visible content with clean [semantic HTML](https://organikpi.com/blog/technical-seo/html5-semantic-tags-ai-content-extraction/) headings but without schema. The visible structure still contributes to AI extraction without creating a type conflict. ## Validation and ongoing monitoring Run every schema implementation through two tools. The Google Rich Results Test (search.google.com/test/rich-results) checks eligibility for specific Google SERP features. FAQPage pages will show “not eligible for FAQ rich results” after the May 2026 deprecation, which is expected. The Schema.org Validator (validator.schema.org) checks schema correctness independently of Google display rules. Use both: the Rich Results Test catches Google-specific property requirements, the Schema.org Validator catches generic schema malformation. For the format decision (JSON-LD vs Microdata), see our companion post on [JSON-LD vs Microdata](https://organikpi.com/blog/technical-seo/structured-data-jsonld-vs-microdata-ai-search/). All implementations above use JSON-LD. Schema selection also interacts with your entity structure: the [sameAs property](https://organikpi.com/blog/technical-seo/schema-sameas-entity-disambiguation-ai-citations/) disambiguates your organization as a citation source, [SpeakableSpecification](https://organikpi.com/blog/technical-seo/speakable-schema-voice-ai-seo/) pairs well with FAQ content for voice AI, and [Review schema](https://organikpi.com/blog/technical-seo/review-schema-g2-trustpilot-ai-brand-citations/) alongside [Person schema](https://organikpi.com/blog/brand-authority/person-schema-author-eeat/) reinforces the E-E-A-T layer AI engines use for citation decisions. Our [GEO audit](https://organikpi.com/services/geo-audit/) reviews all of these layers together. ## Frequently Asked Questions ### Did Google fully remove FAQ rich results and when? Google first restricted FAQ rich results to government and health sites in August 2023. On May 7, 2026, Google dropped the FAQ rich result search appearance entirely for all sites, including removing Search Console tracking and rich result testing support. The schema retained AI citation value through both changes. ### Why does FAQPage schema still drive AI citations after the May 2026 deprecation? AI retrieval systems parse FAQPage schema to extract machine-readable Q&A pairs as discrete, citable chunks. This process runs independently of Google's SERP display rules. Each Question entity becomes a separate answer candidate. Our March 2026 study of 42,971 citations found that pages with structured Q&A markup were cited more often per query across all 6 AI platforms we analyzed. ### What content should use HowTo schema instead of FAQPage? HowTo schema is correct only when your content is a genuine procedural sequence where each step must happen in order and out-of-order execution would cause failure. Tips lists, best practices, and considerations are not HowTo content. They are FAQPage or BlogPosting territory. The test: if skipping step 3 prevents step 4 from working, it is HowTo. If not, use FAQPage. ### What are the required JSON-LD properties for FAQPage schema? FAQPage schema requires @context set to https://schema.org, @type set to FAQPage, and a mainEntity array of Question entities. Each Question needs a name property (the question text) and an acceptedAnswer property containing an Answer entity with a text property (the answer text). Schema content must match the visible Q&A section on the page exactly. ### Can a page use both FAQPage and HowTo schema at the same time? No. Having two competing primary schema types on one page creates entity ambiguity for AI retrieval systems and Google's crawler. The page often gets parsed as neither and loses structured data coverage. If a page has both a procedural section and a Q&A section, pick the dominant type as the schema and render the other section as visible semantic HTML without schema markup.