--- title: "Schema Markup for AEO: A Practical Guide for 2026" description: "Practical guide to schema markup for AEO. Working JSON-LD examples for Organisation, Article, FAQPage, HowTo, Person, Speakable — copy-paste code." image: https://www.mo.agency/hubfs/MO%20-%20Blogs%202026%20-%20Schema%20Markup%20for%20AEO%20-%20A%20Practical%20Guide.png canonical: https://www.mo.agency/blog/schema-markup-for-aeo url: https://ai.mo.agency/blog/schema-markup-for-aeo.md last_converted: 2026-05-11T16:11:06.788Z --- [AEO](https://www.mo.agency/blog/topic/aeo) # Schema Markup for AEO: A Practical Guide for 2026 May 11, 2026 · ![Luke Marthinusen](https://www.mo.agency/hs-fs/hubfs/MO%20-%20New%20Profile%20Picture%20Designs%20-%20Luke%20-%2020240528.png?width=36&height=36&name=MO%20-%20New%20Profile%20Picture%20Designs%20-%20Luke%20-%2020240528.png) Luke Marthinusen ![Schema Markup for AEO](https://www.mo.agency/hs-fs/hubfs/MO%20-%20Blogs%202026%20-%20Schema%20Markup%20for%20AEO%20-%20A%20Practical%20Guide.png?width=1200&height=600&name=MO%20-%20Blogs%202026%20-%20Schema%20Markup%20for%20AEO%20-%20A%20Practical%20Guide.png) Share Most marketers know schema markup as an SEO enhancement - the thing that gets you star ratings in Google or a recipe carousel. That framing is now incomplete. Schema is no longer just structured data for search engines; it is the primary machine-readable layer that AI engines use to understand who you are, what you do, and why they should cite you. This is a practical guide, written for marketers and developers who need working code, not theory. Every schema type below comes with a copy-paste JSON-LD example, where to put it, and what it actually does for Answer Engine Optimisation. If you are new to AEO as a discipline, start with our [What Is Answer Engine Optimisation guide](https://www.mo.agency/blog/what-is-answer-engine-optimisation) and come back here when you are ready to implement. ## Why schema matters more for AEO than it ever did for SEO Search engines use schema as a *signal* - one of many inputs into a ranking algorithm. AI engines use schema as a *source*. That is the shift, and it is significant. When ChatGPT, Claude, Gemini, or Perplexity assemble an answer, they need to extract entities, attributes, and relationships from web content. Unstructured prose forces the model to parse and infer. Structured data gives the model the answer directly. The model does not have to guess that "MO Agency" is an Organisation, that its address is in Johannesburg, that its founder is Luke Marthinusen, or that it offers AEO services - those facts are stated in machine-readable form. The practical implication: pages with comprehensive schema get cited more often, more accurately, and more confidently. In a world where AI Overviews and chat-based answers are eating organic traffic, citation accuracy becomes a survival issue. There is a second-order benefit too. Google publicly [reduced FAQ rich result eligibility in 2023](https://developers.google.com/search/blog/2023/08/howto-faq-changes), which led many SEOs to deprioritise FAQPage schema. AI engines did the opposite - they leaned on it harder, because Q&A-shaped data is exactly what they need for direct-answer generation. If you removed FAQ schema in 2023, put it back. ## The schema types that matter for AEO Not all schema is equally useful for AI citation. The high-leverage types, in priority order: 1. **Organisation** - the foundation. Defines your brand entity. 2. **Article** (or BlogPosting / NewsArticle) - every editorial page. 3. **FAQPage** - direct-answer extraction. 4. **HowTo** - procedural content extraction. 5. **Person** - author authority and E-E-A-T signals. 6. **BreadcrumbList** - site structure and topical hierarchy. 7. **Service / Product** - what you sell, with attributes. 8. **WebSite** - sitelinks and search action. 9. **Speakable** - explicitly designed for voice and answer engines. 10. **ImageObject / VideoObject** - multimedia entity context. The rest of this guide works through each, with code. ## A note on format - JSON-LD only Schema.org supports three serialisation formats: JSON-LD, Microdata, and RDFa. **Use JSON-LD.** It is what Google recommends, what AI crawlers parse most reliably, and what every example in this guide uses. JSON-LD lives in a `` tag in your `` (or ``, but `` is cleaner), separated from your HTML - which means you can add, edit, or remove it without touching content. Here is the basic shape: ``` { "@context": "https://schema.org", "@type": "Organization", "name": "MO Agency" } ``` Every example below assumes this wrapper. ## Organisation schema - your foundation Every site needs Organisation schema sitewide. It defines who you are as an entity, and it is the single most-cited schema type by AI engines when they answer "who is X" or "what does X do" prompts. Place it in the `` of every page, or - better - in the global footer template so it appears on every URL. ``` { "@context": "https://schema.org", "@type": "Organization", "@id": "https://www.mo.agency/#organization", "name": "MO Agency", "url": "https://www.mo.agency", "logo": { "@type": "ImageObject", "url": "https://www.mo.agency/logo.png", "width": 600, "height": 60 }, "description": "MO Agency is a HubSpot Elite Partner specialising in revenue operations, CRM implementations, and digital transformation for enterprise and growth-led companies.", "foundingDate": "2010", "founders": [ { "@type": "Person", "name": "Luke Marthinusen" } ], "address": { "@type": "PostalAddress", "streetAddress": "Hyde Park", "addressLocality": "Johannesburg", "addressRegion": "Gauteng", "postalCode": "2196", "addressCountry": "ZA" }, "contactPoint": { "@type": "ContactPoint", "telephone": "+27-11-XXX-XXXX", "contactType": "Sales", "areaServed": ["ZA", "GB", "US"], "availableLanguage": ["English"] }, "sameAs": [ "https://www.linkedin.com/company/2258363", "https://www.facebook.com/MOSouthAfrica", "https://www.instagram.com/moagencylife" ] } ``` A few things worth noting. The `@id` field is critical - it is the canonical identifier AI engines use to link this entity across pages. Use the same `@id` everywhere you reference your organisation. The `sameAs` array is how you connect your entity to its presence on other authoritative platforms; LinkedIn, Wikipedia, Crunchbase, and Wikidata are the heaviest hitters for AI citation. The `address` block - even with a single physical location - gives AI engines geographic grounding, which is exactly the signal you need when regional intent matters. ## Article schema - every editorial page Every blog post, guide, news article, and long-form page needs Article schema. It tells AI engines what the page is, who wrote it, when it was published, and when it was last updated. Use `BlogPosting` for blog posts, `NewsArticle` for news, and the generic `Article` for everything else. ``` { "@context": "https://schema.org", "@type": "BlogPosting", "@id": "https://www.mo.agency/blog/schema-markup-for-aeo#article", "headline": "Schema Markup for AEO — A Practical Guide", "description": "A practical guide to schema markup for Answer Engine Optimisation, with working JSON-LD examples for the schema types AI engines cite most.", "image": [ "https://www.mo.agency/blog/schema-markup-for-aeo/hero-1200x630.png" ], "datePublished": "2026-04-26T08:00:00+02:00", "dateModified": "2026-04-26T08:00:00+02:00", "author": { "@type": "Person", "@id": "https://www.mo.agency/about#luke-marthinusen", "name": "Luke Marthinusen", "url": "https://www.mo.agency/about", "jobTitle": "CEO", "worksFor": { "@id": "https://www.mo.agency/#organization" } }, "publisher": { "@id": "https://www.mo.agency/#organization" }, "mainEntityOfPage": { "@type": "WebPage", "@id": "https://www.mo.agency/blog/schema-markup-for-aeo" }, "wordCount": 3200, "articleSection": "AEO", "keywords": ["schema markup", "AEO", "answer engine optimisation", "JSON-LD", "structured data"] } ``` Three details that disproportionately affect AEO performance. First, `dateModified` is weighted heavily for retrieval recency - update it every time you genuinely change the page, not just to game freshness. Second, the `author` block must reference a real Person entity (covered below) for E-E-A-T. Third, `wordCount` is a small but real signal of content depth - include it. ## FAQPage schema - direct-answer extraction This is the schema type with the highest AEO leverage relative to effort. AI engines extract Q&A pairs directly from FAQPage markup and serve them as answers, often verbatim. If your page has questions and answers, mark them up. ``` { "@context": "https://schema.org", "@type": "FAQPage", "mainEntity": [ { "@type": "Question", "name": "What is schema markup?", "acceptedAnswer": { "@type": "Answer", "text": "Schema markup is a vocabulary of tags, defined at schema.org, that you add to your website to help search engines and AI engines understand the content. It is the most common machine-readable layer on the web and is implemented as JSON-LD in modern sites." } }, { "@type": "Question", "name": "Does schema markup help with AEO?", "acceptedAnswer": { "@type": "Answer", "text": "Yes. AI engines like ChatGPT, Claude, Gemini, and Perplexity use schema as a primary source of structured information about entities, content, and relationships. Pages with comprehensive schema are cited more often and more accurately than pages without it." } }, { "@type": "Question", "name": "Is FAQPage schema still worth implementing in 2026?", "acceptedAnswer": { "@type": "Answer", "text": "Yes — but for different reasons than it was in 2022. Google reduced FAQ rich results in 2023, but AI engines have leaned on FAQPage schema harder because question-answer pairs are exactly what they need for direct-answer generation. The traffic source has shifted from SERP rich results to AI citations." } } ] } ``` Two practical rules. Keep each answer between 50 and 100 words - long enough to be substantive, short enough to be lifted whole. And do not mark up content that does not exist on the page; the `text` in `acceptedAnswer` should match the actual rendered text. Mismatches are penalised. ## HowTo schema - procedural content If your page describes a process - implementing schema, configuring HubSpot, optimising a workflow - use HowTo schema. AI engines love procedural content because users frequently ask how-to questions, and clean step-by-step extraction is a guaranteed citation winner. ``` { "@context": "https://schema.org", "@type": "HowTo", "name": "How to implement Organisation schema on your website", "description": "A four-step process to add Organisation schema markup to a website so that AI engines and search engines can recognise your brand entity.", "totalTime": "PT30M", "step": [ { "@type": "HowToStep", "position": 1, "name": "Gather your organisation details", "text": "Collect the canonical version of your organisation name, logo URL, address, founding date, founder names, and social profile URLs. These are the inputs to the schema." }, { "@type": "HowToStep", "position": 2, "name": "Build the JSON-LD", "text": "Use the Organisation schema template with your details. Include @id, name, url, logo, description, address, contactPoint, and sameAs at minimum." }, { "@type": "HowToStep", "position": 3, "name": "Place the schema in your global template", "text": "Add the JSON-LD inside a script tag with type='application/ld+json' in the head section of your global site template, so it appears on every page." }, { "@type": "HowToStep", "position": 4, "name": "Validate with Google's Rich Results Test and Schema.org Validator", "text": "Run the page through both validators. Fix any errors and warnings. Re-validate after every meaningful change." } ] } ``` A note on `totalTime` - use ISO 8601 duration format. `PT30M` is 30 minutes, `PT1H30M` is one hour thirty minutes, `P1D` is one day. Get this right. ## Person schema - author authority AI engines weight author authority heavily, especially Claude and Perplexity. A page with a properly marked-up author who has clear credentials, a job title, and external authority signals (LinkedIn, published work) outperforms a brand-only byline by a meaningful margin. ``` { "@context": "https://schema.org", "@type": "Person", "@id": "https://www.mo.agency/about#luke-marthinusen", "name": "Luke Marthinusen", "url": "https://www.mo.agency/about", "image": "https://www.mo.agency/team/luke-marthinusen.jpg", "jobTitle": "CEO", "worksFor": { "@id": "https://www.mo.agency/#organization" }, "knowsAbout": [ "HubSpot", "Revenue Operations", "Answer Engine Optimisation", "Digital Transformation", "CRM Implementation" ], "alumniOf": { "@type": "EducationalOrganization", "name": "University of the Witwatersrand" }, "sameAs": [ "https://www.linkedin.com/in/lukemarthinusen", "https://twitter.com/lukemarthinusen" ] } ``` The `knowsAbout` array is genuinely useful - it tells AI engines what topics you are an authority on, which is exactly the signal that gets you cited when those topics come up. Be specific. "Marketing" is too broad; "Answer Engine Optimisation" is precise enough to be a citation handle. ## BreadcrumbList schema - site structure Breadcrumbs give AI engines topical hierarchy - they show how a page sits within your site's content tree. This matters because models use hierarchy to assess depth and specialism. A page about schema markup nested under `Blog > AEO > Technical Implementation` carries more topical weight than the same page sitting at the root. ``` { "@context": "https://schema.org", "@type": "BreadcrumbList", "itemListElement": [ { "@type": "ListItem", "position": 1, "name": "Home", "item": "https://www.mo.agency" }, { "@type": "ListItem", "position": 2, "name": "Blog", "item": "https://www.mo.agency/blog" }, { "@type": "ListItem", "position": 3, "name": "AEO", "item": "https://www.mo.agency/blog/category/aeo" }, { "@type": "ListItem", "position": 4, "name": "Schema Markup for AEO" } ] } ``` Note that the final item - the current page - has no `item` URL. That is intentional and correct. ## Service schema - what you sell For service pages, use Service schema (or Product if you sell products). This is what tells AI engines what you actually offer, which is the schema type they consult when answering vendor-recommendation prompts. ``` { "@context": "https://schema.org", "@type": "Service", "@id": "https://www.mo.agency/solutions/demand-generation/aeo-services#service", "name": "Answer Engine Optimisation (AEO) Services", "description": "AEO services that get your brand cited and recommended by ChatGPT, Claude, Gemini, and Perplexity. Includes AI visibility auditing, entity optimisation, structured data implementation, content strategy, and citation building.", "provider": { "@id": "https://www.mo.agency/#organization" }, "serviceType": "Answer Engine Optimisation", "areaServed": [ { "@type": "Country", "name": "South Africa" }, { "@type": "Country", "name": "United Kingdom" }, { "@type": "Country", "name": "United States" } ], "hasOfferCatalog": { "@type": "OfferCatalog", "name": "AEO Service Tiers", "itemListElement": [ { "@type": "Offer", "name": "AI Visibility Audit", "description": "Initial benchmarking of brand presence across ChatGPT, Claude, Gemini, and Perplexity.", "priceCurrency": "USD", "price": "2500" }, { "@type": "Offer", "name": "AEO Monthly Retainer", "description": "Ongoing AEO programme including content strategy, structured data, citation building, and monthly measurement.", "priceCurrency": "USD", "price": "2500", "priceSpecification": { "@type": "UnitPriceSpecification", "price": "2500", "priceCurrency": "USD", "unitText": "MONTH" } } ] } } ``` The `hasOfferCatalog` block is what gets you into vendor-comparison answers when an AI engine is asked "what does X cost" or "compare AEO providers". Put real pricing here if you can - vagueness gets discounted. ## WebSite schema - site-level entity WebSite schema is added once, sitewide, and tells AI engines about your site as an entity rather than your organisation as an entity. It also enables the SearchAction box that some AI engines now surface. ``` { "@context": "https://schema.org", "@type": "WebSite", "@id": "https://www.mo.agency/#website", "url": "https://www.mo.agency", "name": "MO Agency", "description": "HubSpot Elite Partner specialising in RevOps, AEO, and digital transformation for enterprise and growth-led companies.", "publisher": { "@id": "https://www.mo.agency/#organization" }, "potentialAction": { "@type": "SearchAction", "target": { "@type": "EntryPoint", "urlTemplate": "https://www.mo.agency/search?q={search_term_string}" }, "query-input": "required name=search_term_string" }, "inLanguage": "en-GB" } ``` ## Speakable schema - explicitly built for answer engines Speakable schema is the most under-used type in AEO right now. It explicitly marks sections of a page as suitable for voice and audio rendering - which is precisely how many AI assistants surface content. Google's documentation labels it experimental, which has scared marketers off, but Perplexity, voice assistants, and increasingly Gemini all factor it in. ``` { "@context": "https://schema.org", "@type": "Article", "@id": "https://www.mo.agency/blog/schema-markup-for-aeo#article", "headline": "Schema Markup for AEO — A Practical Guide", "speakable": { "@type": "SpeakableSpecification", "cssSelector": [".article-summary", ".key-takeaways", ".faq-section"] } } ``` The `cssSelector` array points to elements on the page that contain answer-shaped content. For best results, give your page a `.article-summary` block at the top, a `.key-takeaways` block somewhere prominent, and a `.faq-section` block at the bottom. This gives AI engines three discrete, well-bounded sections to extract. ## Where to put schema in your site Three placement principles: 1. **Sitewide schema in the global template.** Organisation, WebSite, and the main navigation BreadcrumbList go in your header or footer template, so they appear on every URL. 2. **Page-specific schema in the page template.** Article, FAQPage, HowTo, Service - these go in templates for the specific page type. 3. **Bundle related schema in a single script tag where possible.** A blog post page can have one JSON-LD block containing Article + Person + BreadcrumbList + FAQPage. Multiple separate `` tags are fine, but a `@graph` array is cleaner: ``` { "@context": "https://schema.org", "@graph": [ { "@type": "BlogPosting", "@id": "https://www.mo.agency/blog/schema-markup-for-aeo#article", "headline": "Schema Markup for AEO — A Practical Guide" }, { "@type": "Person", "@id": "https://www.mo.agency/about#luke-marthinusen", "name": "Luke Marthinusen" }, { "@type": "BreadcrumbList", "itemListElement": [] } ] } ``` ## Validation - non-optional Every page with schema markup must be validated. Two tools, both free: 1. [Schema.org Validator](https://validator.schema.org) - checks against the schema.org spec. Use this first. 2. [Google Rich Results Test](https://search.google.com/test/rich-results) - checks against Google's specific implementation, which is stricter. Use this second. For monitoring at scale, [Ahrefs Site Audit](https://ahrefs.com/site-audit) flags schema errors across an entire site and is what we use for client work. A schema implementation with errors is sometimes worse than no schema at all - broken markup can fragment the entity in AI engines' understanding. Validate every change, every time. ## Common schema mistakes Things we see frequently in audits: - **Marking up content that is not on the page.** The schema must reflect rendered HTML. Phantom FAQ items, made-up reviews, fictional pricing - all penalised. - **Inconsistent @id values across pages.** The same Organisation must have the same `@id` everywhere it appears. Inconsistent IDs fragment the entity. - **Skipping sameAs.** This is the array that links your entity to LinkedIn, Wikipedia, and Wikidata - the platforms AI engines weight most heavily for verification. - **Outdated dateModified.** If you are not updating this when you genuinely revise content, you are leaving recency signal on the table. - **Author as a string, not a Person object.** `"author": "Luke Marthinusen"` is technically valid but provides no authority signal. Use the full Person object. - **Forgetting BreadcrumbList.** Easy win, frequently skipped. - **No mainEntityOfPage on Article schema.** Without this, models can struggle to bind the article to the URL it appears on. - **HTML in JSON-LD strings.** All values must be plain text. No ``, no ``, no entities. Strip all HTML before serialising. ## Schema and llms.txt - the complete machine-readable layer Schema is one half of the machine-readable web. The other half is [llms.txt](https://llmstxt.org), an emerging standard that gives AI crawlers a clean, navigable index of your most important content. Together, schema and llms.txt form the layer AI engines use to understand and cite a site. If you are already implementing schema seriously, llms.txt is a small additional lift with disproportionate upside. The fastest way to deploy both - and to monitor which AI crawlers are actually consuming them - is [getMD.ai](https://getmd.ai), which serves machine-readable markdown versions of every page on your site and logs AI crawler activity in real time. ## What every AEO-focused page should have If you implement nothing else, implement this stack: | Page type | Required schema | | --- | --- | | Every page (sitewide) | Organisation, WebSite, BreadcrumbList | | Blog posts | BlogPosting + Person (author) + FAQPage (if applicable) | | Service pages | Service + Offer (with pricing) + FAQPage | | Product pages | Product + Offer + AggregateRating + Review | | Case studies | Article + Person (client contact, if approved) + Organisation (client) | | About / team pages | Person (each team member) + Organisation | | Procedural / how-to content | HowTo + FAQPage | That is the floor. Everything beyond it - Speakable, ImageObject, VideoObject, Event, Course - is upside. ## Frequently asked questions ### What is schema markup? Schema markup is a vocabulary of tags defined at schema.org that you add to your website to help search engines and AI engines understand your content. It is implemented as JSON-LD on modern sites and acts as the primary machine-readable layer that AI engines parse when assembling answers and citations. ### Does schema markup help with AEO? Yes. AI engines like ChatGPT, Claude, Gemini, and Perplexity use schema as a primary source of structured information. Pages with comprehensive, valid schema are cited more often and more accurately than pages without it. The leverage is highest on Article, FAQPage, HowTo, Organisation, and Person schema. ### Is FAQPage schema still worth implementing in 2026? Yes, more than ever - but for different reasons than in 2022. Google reduced FAQ rich results in 2023, but AI engines have leaned on FAQPage schema harder because question-answer pairs are exactly what they need for direct-answer generation. The traffic source has shifted from SERP rich results to AI citations. ### What is the difference between schema and JSON-LD? Schema (or schema.org) is the vocabulary - the list of available types and properties. JSON-LD is the format used to express that vocabulary in code. You implement schema *using* JSON-LD. The other formats - Microdata and RDFa - exist but are largely deprecated; use JSON-LD only. ### How long does it take to see AEO results from schema? Initial AI engine indexing of new schema typically takes two to six weeks. Citation behaviour starts shifting four to twelve weeks after that, depending on the engine. Perplexity moves fastest because it retrieves at query time. ChatGPT moves slowest because of training-data cycles. Google AI Overviews track Google's normal index, so changes there are reflected in days to weeks. ### Can I use multiple schema types on one page? Yes, and you should. Bundle related schema in a single JSON-LD block using the `@graph` array, or use multiple `` tags - both are valid. A blog post commonly carries Article + Person + BreadcrumbList + FAQPage in one combined graph. ### Where should I put my JSON-LD? In the `` of the page where possible, inside a `` tag. Sitewide schema (Organisation, WebSite) goes in your global template. Page-specific schema goes in the page-type template. ### Do I need a developer to add schema? For a one-off page, no - most CMS platforms have schema plugins that handle the heavy lifting. For a serious AEO programme across a site of any size, yes - you want a developer to implement schema as part of the page templates, not as bolt-on plugins. Plugins are inconsistent and frequently produce malformed output. ### How do I validate my schema? Use the [Schema.org Validator](https://validator.schema.org) for spec compliance and the [Google Rich Results Test](https://search.google.com/test/rich-results) for Google's stricter implementation rules. Validate every change, every time. ### What is Speakable schema and should I use it? Speakable schema marks sections of a page as suitable for voice and audio rendering. It is officially labelled experimental by Google but is increasingly factored in by AI engines and voice assistants. The cost to implement is low. The upside is real. Use it. ## Where to next Schema is one of the four pillars of Answer Engine Optimisation, alongside authoritative content, citation building, and AI visibility measurement. To go deeper: - **[What Is Answer Engine Optimisation? The 2026 Guide](https://www.mo.agency/blog/what-is-answer-engine-optimisation)** - the foundational guide to AEO as a discipline. - **[AEO Services](https://www.mo.agency/solutions/demand-generation/aeo-services)** - how MO Agency implements AEO programmes for enterprise and growth-led companies. - **[Getmd.ai](https://getmd.ai)** - the fastest way to make your site machine-readable for AI engines and to monitor which crawlers are reading it. If you would like a free AI Visibility Audit - including a schema implementation review - [book a discovery call](https://www.mo.agency/contact). We will benchmark your current presence across ChatGPT, Claude, Gemini, and Perplexity and show you exactly where the schema gaps are. [Scroll to top](#top-banner)