Schema markup matters more for AI search than it ever did for Google. Not because AI assistants display rich snippets — they don't — but because structured data gives the retriever clearer anchors and the language model cleaner extraction targets.

This post is the practical guide we use internally at Tracemetry to decide what schema to put on which page.

Why schema matters more now

Google rewards schema with rich results — stars, FAQ accordions, breadcrumbs. The benefit is mostly cosmetic.

AI assistants don't display schema visually. The benefit is upstream: schema helps the retriever and the language model understand what a page contains. A Product block says "this page is a product, here's the name, price, availability, rating." A ComparisonPage block says "this page compares specific items along specific dimensions."

When a language model has to extract a fact from your page, schema gives it explicit anchors. That meaningfully increases your citation rate.

The five schema types that matter for AI search

1. Article

The default for blog posts, guides, and editorial content.

{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "How to rank in ChatGPT answers",
  "description": "What it takes to be the brand ChatGPT recommends.",
  "author": {
    "@type": "Person",
    "name": "Jane Doe",
    "url": "https://yoursite.com/authors/jane"
  },
  "datePublished": "2026-05-14",
  "dateModified": "2026-05-14",
  "publisher": {
    "@type": "Organization",
    "name": "Tracemetry",
    "logo": { "@type": "ImageObject", "url": "https://tracemetry.com/brand/icon-512.png" }
  },
  "mainEntityOfPage": {
    "@type": "WebPage",
    "@id": "https://tracemetry.com/blog/how-to-rank-in-chatgpt"
  }
}

What matters most:

• author — a real person, not "Editorial Team." Link to an author page if possible.
• datePublished + dateModified — accurate, not faked.
• publisher — your org with a logo URL.
• mainEntityOfPage — the canonical URL of the page.

2. FAQPage

For any page with a Q&A section, this is the highest-ROI schema type for AI search.

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "How often does Tracemetry re-check prompts?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Weekly on Tracker and Pro plans. Daily on Agency. On-demand from the dashboard."
      }
    }
  ]
}

Why it works for AI search: the question-answer structure exactly matches how AI assistants parse content. The retriever sees explicit Q→A pairs and the language model lifts them cleanly.

Put FAQPage schema on:

• Pricing pages with FAQs
• Product pages with FAQs
• Help articles
• Any blog post with a Q&A section

3. ComparisonPage / Product

For comparison content and product pages.

ComparisonPage is technically an experimental schema type (it isn't in core schema.org), so most teams use Article or Product with explicit additionalType or about fields naming what's being compared.

The simpler and more widely-supported pattern: Product with offers, aggregateRating, and review markup.

{
  "@context": "https://schema.org",
  "@type": "Product",
  "name": "Tracemetry Pro",
  "description": "AI visibility platform: 250 tracked prompts, source-grounded content generation.",
  "brand": { "@type": "Brand", "name": "Tracemetry" },
  "offers": {
    "@type": "Offer",
    "price": "199",
    "priceCurrency": "USD",
    "url": "https://tracemetry.com/pricing"
  }
}

Why it matters: AI assistants get a lot of "best X for Y" prompts. Pages with clean Product schema get pulled into shortlist answers far more often than pages without.

4. HowTo

For step-by-step content.

{
  "@context": "https://schema.org",
  "@type": "HowTo",
  "name": "How to track your AI visibility weekly",
  "step": [
    {
      "@type": "HowToStep",
      "name": "Pick 25 category-relevant prompts",
      "text": "Identify the prompts your buyers ask AI assistants."
    },
    {
      "@type": "HowToStep",
      "name": "Run each through ChatGPT, Claude, and Perplexity",
      "text": "Record whether your brand is mentioned and which competitors appear."
    }
  ]
}

Why it matters: AI assistants love step-by-step content. HowTo schema makes the steps machine-readable, which dramatically increases the likelihood the page gets extracted for "how do I X" prompts.

5. Organization

Top-level identity schema. Goes on every page (usually via the root layout).

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "Tracemetry",
  "url": "https://tracemetry.com",
  "logo": "https://tracemetry.com/brand/icon-512.png",
  "description": "AI visibility platform...",
  "contactPoint": {
    "@type": "ContactPoint",
    "contactType": "customer support",
    "email": "support@tracemetry.com"
  }
}

Why it matters: gives AI assistants a clean entity to reference when surfacing your brand. Also feeds into Google Knowledge Panel and similar entity graphs.

Schema patterns that don't help (and may hurt)

A few things to avoid:

• Review schema with fake or self-authored reviews. AI assistants and Google both penalize this.
• AggregateRating with no real data behind it. Same.
• Article with bogus author blocks (e.g., "Editorial Staff" as a Person). Use Organization if it's not a real person.
• Over-decorated schema with twenty fields, half of which are placeholders. Better to ship minimal correct schema than maximal wrong schema.
• Conflicting schema blocks on the same page. If you have a Product and an Article block both claiming mainEntityOfPage, pick one.

How to audit your existing schema

A three-step audit for any site:

1. Google Rich Results Test

Google's Rich Results Test is the simplest validator. Paste a URL, see what schema Google detects, see what passes validation.

Pages that should have schema but don't show any: top priority to add.

Pages with schema that has errors: fix the errors before adding more.

2. The "could AI cite this?" test

For each of your top 20 pages by traffic:

• Is there an Article block with real author and datePublished?
• Is there FAQPage schema where the page has FAQs?
• Is there Product schema where the page is a product?
• Is there HowTo schema where the page is step-by-step?

If you fail more than 5 of these across 20 pages, schema is your cheapest GEO win.

3. The 1-week shipping audit

Pick the 10 highest-traffic pages on your site. Add or fix the appropriate schema for each one in a single week. Then watch your AI visibility for the next 4 weeks.

We've seen this exact playbook move presence 5–15 points on its own — and it's pure infrastructure work, no new content needed.

Tracemetry's approach

When we generate drafts for our customers, the brief includes a schema_recommendation field — a specific list of schema types the page should ship with. The draft engine produces those blocks inline. The publisher pushes them through to your CMS.

You don't have to think about this every time you ship. The brief tells the writer (or the generator) what schema to include, and the published page ships with it.

If you want to see what schema Tracemetry would recommend for your top opportunities:

• Free public audit — submit your domain, we'll surface 3 specific opportunities
• Get started on Pro — full prompt universe + briefs with schema recommendations

Schema is the cheapest, most under-used GEO lever. Ship the right blocks on your top 10 pages and you'll see the lift.