When ChatGPT, Perplexity, or Google's AI Overview assembles an answer about your industry, it doesn't read your website the way a human does. It parses HTML, extracts facts, weighs confidence, and picks sources. If your site doesn't speak that language clearly, you get skipped — even if you're the best option in your market.
Schema markup is how you speak that language. It's structured data you embed in your HTML that tells AI systems, search engines, and language models exactly what your company is, what you offer, and why you can be trusted. This tutorial walks you through the exact implementation, page by page, with copy-ready code.
What this tutorial covers: the format that works for AI (JSON-LD), the five schema types that matter most for B2B companies, ready-to-use code blocks for each, validation steps, and how to confirm that AI tools are actually reading your markup.
Why AI tools need structured data (and why HTML alone isn't enough)
AI crawlers — including GPTBot (ChatGPT), ClaudeBot (Claude), and PerplexityBot — fetch your raw HTML and stop. They don't execute JavaScript. They don't wait for client-side rendering. If your content loads dynamically, those crawlers see a blank page regardless of how beautifully your site renders in a browser.
Even when they do see your content, they have to guess at what things mean. "AquaFlow Systems" on a page could be your company name, a product line, or a competitor you mentioned. Structured data removes the guessing. It tells the machine: this is an Organization, here is its name, here is what it does, here is who it serves.
Pages with complete Organization, Service, and Article schema appear in ChatGPT citations 3.4 times more often than pages with only basic Open Graph tags.
The implication for niche B2B companies is significant. If your competitors in a small market haven't implemented schema, you can take a meaningful lead simply by making your site machine-readable.
Use JSON-LD — not Microdata or RDFa
There are three ways to add structured data to a page: JSON-LD, Microdata, and RDFa. For AI visibility purposes, use JSON-LD exclusively. Here's why:
- JSON-LD lives in a
<script>tag in your page's<head>, completely separate from your visible HTML. AI crawlers can extract it cleanly without parsing your layout. - Google explicitly recommends JSON-LD as the preferred format for structured data.
- It's the easiest to add, update, and validate without touching your content.
Every code block in this tutorial uses JSON-LD. Place each one inside a <script type="application/ld+json"> tag in the <head> section of the relevant page.
The five schema types every B2B site needs
Who you are
Your company's identity, contact info, and social profiles. Goes on your homepage and About page.
What you do
Describes each offering with a name, description, provider, and area served. Goes on each service page.
Your editorial content
Marks up blog posts and guides with headline, author, dates, and topic. Goes on every blog post.
Direct answers
Wraps question-and-answer sections so AI tools can extract them verbatim. Goes anywhere you have FAQs.
Site structure
Tells AI systems where each page sits in your site hierarchy. Goes on every page that has a breadcrumb.
Step-by-step content
Marks up tutorials and process guides. AI tools pull steps from HowTo schema directly into answers.
Step 1: Organization schema on your homepage
This is the most important piece of schema on your entire site. It establishes your company as a named entity in the knowledge graph. Without it, AI tools may confuse you with similarly named businesses, misstate your services, or simply not know enough about you to include you in an answer.
Add this to your homepage's <head> section. Replace the placeholder values with your real information:
<script type="application/ld+json"> { "@context": "https://schema.org", "@type": "Organization", "name": "Acme Industrial Filtration", "url": "https://yourcompany.com", "logo": "https://yourcompany.com/logo.png", "description": "Industrial water filtration systems for mid-size manufacturers in the US and Canada. Specialising in food-grade and pharmaceutical applications.", "foundingDate": "2008", "areaServed": ["US", "CA"], "contactPoint": { "@type": "ContactPoint", "telephone": "+1-800-555-0100", "contactType": "sales" }, "sameAs": [ "https://linkedin.com/company/acme-filtration", "https://twitter.com/acmefiltration" ] } </script>
Key field to get right: the description. Write it in one to two sentences that clearly state your niche, the customer type you serve, and the geography or industry you cover. This is the sentence AI tools quote when they describe your company.
Step 2: Service schema on each service page
If a buyer asks ChatGPT "who provides [your specific service type] for [your industry]," your service pages are where the answer gets built. Give each one its own Service schema block.
<script type="application/ld+json"> { "@context": "https://schema.org", "@type": "Service", "name": "Pharmaceutical-Grade Water Filtration Systems", "description": "Custom water filtration systems built to FDA 21 CFR Part 11 standards for pharmaceutical manufacturers. Includes installation, validation, and ongoing maintenance.", "provider": { "@type": "Organization", "name": "Acme Industrial Filtration", "url": "https://yourcompany.com" }, "areaServed": "US", "serviceType": "Industrial Filtration", "audience": { "@type": "Audience", "audienceType": "Pharmaceutical Manufacturers" } } </script>
Step 3: Article schema on every blog post
Blog posts are your main citation surface. An AI tool deciding whether to quote your content weighs freshness, authorship, and topical clarity. Article schema provides all three explicitly.
The two fields most people miss: dateModified (update this every time you revise the post — AI tools use it to assess freshness) and about (a list of topics the article covers, which helps the AI match your post to relevant queries).
<script type="application/ld+json"> { "@context": "https://schema.org", "@type": "Article", "headline": "How to Choose a Water Filtration System for Food Processing", "description": "A buyer's guide to selecting industrial water filtration for food and beverage manufacturing, including capacity, certification, and maintenance requirements.", "author": { "@type": "Organization", "name": "Acme Industrial Filtration" }, "publisher": { "@type": "Organization", "name": "Acme Industrial Filtration" }, "datePublished": "2026-03-15", "dateModified": "2026-06-20", "mainEntityOfPage": "https://yourcompany.com/blog/water-filtration-food-processing", "about": [ { "@type": "Thing", "name": "Industrial Water Filtration" }, { "@type": "Thing", "name": "Food Processing Equipment" } ] } </script>
Step 4: FAQPage schema on any page with Q&A content
FAQPage schema is among the highest-leverage implementations for AI citation. When someone asks a language model a question, it looks for pre-packaged answer blocks. FAQPage schema is exactly that — a structured list of questions with complete, self-contained answers.
Write answers that stand alone without context. Each one should be a complete statement that makes sense even without reading the surrounding page.
<script type="application/ld+json"> { "@context": "https://schema.org", "@type": "FAQPage", "mainEntity": [ { "@type": "Question", "name": "What size filtration system do I need for a 50,000 sq ft manufacturing facility?", "acceptedAnswer": { "@type": "Answer", "text": "A 50,000 sq ft food processing facility typically requires a filtration system rated for 20 to 40 gallons per minute, depending on production volume and water quality requirements. We recommend a site assessment before specifying capacity." } }, { "@type": "Question", "name": "How often does an industrial water filter need to be serviced?", "acceptedAnswer": { "@type": "Answer", "text": "Most industrial water filtration systems require a preventive maintenance visit every 3 to 6 months. High-throughput food processing facilities may require quarterly service to maintain NSF certification compliance." } } ] } </script>
Important: Every question and answer in your FAQPage schema must match content that is visible on the page. If an AI system fetches your page and sees schema claiming you answered a question that doesn't appear in the HTML, it will reduce its confidence in your entire site.
Step 5: Add Article schema to every blog post (a common miss)
It's extremely common for B2B companies to publish genuinely useful content that never gets cited by AI tools — simply because the blog posts have no schema at all. Article schema takes five minutes to add per post and meaningfully increases your citation rate.
If you're on WordPress, RankMath and Yoast both generate Article schema automatically. On other platforms, paste the JSON-LD block into the page's <head>. After adding it, use Google's Rich Results Test to confirm the markup is valid before the post goes live.
Step 6: Validate before you ship
Invalid schema is worse than no schema. An error in your JSON-LD can cause the entire block to be ignored, and in some cases can flag your domain for inconsistent structured data. Always validate with these two tools:
- Google Rich Results Test (search.google.com/test/rich-results) — paste your URL or your raw schema code and check for errors and warnings.
- Schema.org Validator (validator.schema.org) — stricter conformance check against the Schema.org specification.
Step 7: Confirm that AI tools are actually reading it
After your schema is live and validated, test it directly in the tools you're optimizing for. Go to ChatGPT or Perplexity and ask: "What does [your company name] do?" or "Who are the main providers of [your service] in [your geography]?"
If the AI returns accurate, specific information from your Organization schema — your description, your service types, your area served — the schema is being read. If it returns vague or incorrect information, check that the page is indexable and that the crawlers aren't blocked in your robots.txt.
Add Organization schema to homepage
Name, description, URL, contact, and sameAs social profiles. This is your AI identity document.
Add Service schema to each offering page
Include provider, audience, area served, and a specific service description.
Add Article schema to every blog post
Include headline, datePublished, dateModified, and an about array listing your topics.
Add FAQPage schema wherever you have Q&A
Write answers that are complete standalone statements, not sentence fragments.
Validate with Rich Results Test
No errors, no warnings. Fix them before publishing or updating.
Test in ChatGPT and Perplexity
Ask real buyer questions. If the AI cites your specific facts accurately, the schema is working.
What to avoid
A few common mistakes that can hurt more than help:
- Schema that doesn't match your visible content. If your Article schema says "Published: January 2026" but the page shows a different date, AI systems detect the mismatch and reduce confidence in your domain.
- Overstuffing schema with keywords. AI tools detect this. Write schema descriptions the same way you'd write them for a human reader.
- Forgetting to update dateModified. Every time you revise a blog post, update this field. Freshness is a real signal in AI citation decisions.
- Adding schema to pages that are blocked in robots.txt. If AI crawlers can't access the page, the schema doesn't help.
Want us to audit your schema and fix it?
We check every page on your site for missing, broken, and mismatched schema, then implement the fixes and confirm they're being read by ChatGPT, Claude, and Perplexity.
Request a schema auditFrequently asked questions
What is the most important schema type for B2B AI visibility?
Organization schema is the highest priority because it establishes your business as a distinct entity that AI systems can reliably identify. After that, FAQPage and Service schema give AI the specific facts it needs to cite you in relevant answers.
Does schema markup directly improve my Google ranking?
Google has stated that schema is not a direct ranking factor, but it improves click-through rates through rich results and helps AI systems like Google AI Overviews cite your content. The indirect effect on visibility is significant.
How long does it take to see results from schema markup?
Google typically crawls and validates schema within 1 to 2 weeks. Appearing in AI tool citations can happen within 2 to 6 weeks after implementation, depending on how frequently those tools update their indexes.
Can I add schema markup without a developer?
Yes. JSON-LD schema goes in the HTML head section as a script tag. If your site uses WordPress, plugins like RankMath or Yoast can generate basic schema automatically. For more advanced implementations, a developer will speed things up.