What is schema markup and why does it matter in 2026?
Schema markup is structured data you add to your HTML in JSON-LD format. Search engines read it to understand what a page actually is: a service, an article, a review, a business, a product, an FAQ.
Until 2024 schema was mainly a rich-result feature: review stars in search results, FAQ accordions, product prices. In 2026 the stakes are higher: AI Overviews, ChatGPT Search and Perplexity cite websites based on structured data. No schema means almost no citation.
For SMEs this means you implement structured data not just for Google ranking, but for your entire AI-search visibility in tomorrow's search engines.
Which schema types do you really need?
For 80% of SME sites, 5 schema types are enough: Organization, LocalBusiness, Service, Article and FAQPage. Plus BreadcrumbList as a global addition on every page.
Organization: global identity of your business. Belongs once in your site layout, not per page.
LocalBusiness (or ProfessionalService for services): for anyone with a physical location or service area. Crucial for Local Pack position in Google Maps.
Service: separately per service page, with serviceType, areaServed and provider link.
Article (or BlogPosting for blog content): for substantive articles. Requires an image, author and datePublished for rich result eligibility.
FAQPage: for FAQ blocks. Since 2023 no Google rich result for commercial sites but still crucial for AI citation. ChatGPT and Perplexity cite FAQ answers directly.
Code example: LocalBusiness for SMEs
A valid LocalBusiness schema for a Tilburg SME looks like this:
{ "@context": "https://schema.org", "@type": "ProfessionalService", "name": "Business Name", "telephone": "+31132340371", "email": "info@example.nl", "address": { "@type": "PostalAddress", "addressLocality": "Tilburg", "addressRegion": "Noord-Brabant", "addressCountry": "NL" }, "areaServed": [{"@type": "City", "name": "Tilburg"}, {"@type": "City", "name": "Eindhoven"}], "priceRange": "€€" }
Place it as a JSON-LD script tag in your site layout. A typical Next.js implementation loads this via a schema helper function. WordPress sites do it via an SEO plugin like Yoast or Rank Math.
Code example: FAQPage for AI citation
FAQ schema is the simplest way to land in AI Overviews. Every question needs a direct, citation-ready answer.
{ "@context": "https://schema.org", "@type": "FAQPage", "mainEntity": [{ "@type": "Question", "name": "What does SEO cost for SMEs?", "acceptedAnswer": { "@type": "Answer", "text": "SEO engagements for SMEs start from €600 per month. Exact pricing depends on competition and site scope." } }] }
The answer text must be independently readable — an AI must be able to cite it without context from the rest of the page. Keep answers under 80 words for maximum citation chance.
Validation: how to check if it works
Google Rich Results Test (search.google.com/test/rich-results) shows per URL whether the schema is valid and which rich results are eligible. First stop with every implementation.
Schema.org Validator (validator.schema.org) is stricter and catches subtle errors that Google's tool misses. Use both.
Google Search Console > Enhancements shows per schema type how many pages produce rich results, and which errors or warnings exist.
For AI Overviews: no official validator yet. Check manually by searching your brand name on Perplexity and seeing which pages do or don't get cited.
Common mistakes
Inventing aggregate ratings: schema star reviews that don't exist = Google penalty. Only apply this when you have verifiable reviews.
Using FAQPage for marketing copy: Google has disabled FAQ rich results for commercial sites. You lose nothing by adding FAQ schema, but Google no longer shows it as an accordion. Still crucial for AI citation.
Service schema without areaServed: search engines don't know who your service is available for. Always state your service area.
Confusing Organization vs LocalBusiness: Organization is global (an Org doesn't need a location). LocalBusiness specifically recognises a venue with an address. Use LocalBusiness if you have a service address.
Schema markup and AI Overviews: what changed in 2026?
Google's AI Overviews since Q1 2026 lean heavily on structured data to choose sources. Sites without schema almost never appear as a citation source.
ChatGPT Search (visible to all users by default since January 2026) uses schema to identify entities. A good Organization schema with sameAs to LinkedIn and other profiles raises brand recognition.
Perplexity mainly cites Article schema with dateModified and image. Pages with static dateModified values are prioritised lower. Implement dynamic dateModified (build-time or CMS-edit-time).
Implementation checklist
1. Site-wide Organization in your layout: once per site.
2. LocalBusiness/ProfessionalService on your homepage and contact page.
3. Service schema on each service page.
4. Article (or BlogPosting) on each blog post with image and author.
5. FAQPage on all pages with FAQ blocks.
6. BreadcrumbList on every page except homepage.
7. Validate with Rich Results Test + schema.org Validator.
8. Monitor weekly via GSC > Enhancements for errors.
Want your site implemented schema-correctly for AI-search eligibility?
Request a schema audit →