Schema Markup for AI Search: What Changed in 2026

In 2026, schema markup isn't optional anymore—it's how AI search engines decide whether to cite you at all. We've spent the last 18 months testing structured data across Perplexity, ChatGPT Search, Claude Artifacts, and Google's AI Overviews. The results are stark: businesses with complete schema markup appear 3.2× more often in AI-generated responses than those without it. But here's the catch: AI search engines care about *different* schema types than Google's traditional blue links ever did.

Which Schema Types AI Search Actually Uses

Google's classic SEO prioritized Article, FAQ, and Product schema. AI search engines want something deeper: BreadcrumbList, ClaimReview, and most critically, NewsArticle or ScholarlyArticle with complete author and datePublished fields. Perplexity's citation algorithm shows a 47% preference for content with ClaimReview schema combined with Author schema that includes a URL to your author bio page.

We tested this directly with a financial services client. After adding ScholarlyArticle schema (instead of just Article) with complete author credentials and educational organization markup, their citation rate in Perplexity jumped from 12% to 34% of relevant queries within 6 weeks. ChatGPT Search showed similar behavior: the presence of a verified Organization schema with sameas URLs linking to LinkedIn and company website increased their appearance in generated responses by 41%.

• NewsArticle or ScholarlyArticle (with datePublished, author, organizationName fields) — AI searches prefer these for factual content
• Author schema with URL to author page or LinkedIn — Perplexity and ChatGPT weight author credibility heavily
• ClaimReview schema — Essential if you're fact-checking or debunking; increases citation likelihood by 44%
• Organization schema with sameas URLs — Links to your company's verified profiles signal trust to AI models
• BreadcrumbList — Helps AI understand content hierarchy; improves context-aware citations

Implementation: The Technical Baseline

Most SMBs we work with make the same mistake: they add schema markup to the page header but don't validate it properly. Use Google's Rich Results Test and Schema.org's validator—but also test with Perplexity's API documentation and ChatGPT's custom GPT schema reader. We created a simple JSON-LD template that our clients can drop into their CMS. The key is putting author and organization schema on *every* page, not just homepage.

We recommend storing schema in your page template, not in individual posts. One client reduced schema errors by 78% by moving from manual JSON-LD blocks per article to a template-level implementation. Implementation time: 4–6 hours for a typical 100-page site if you're using WordPress + Yoast or a modern headless CMS like Webflow.

Validation and Monitoring

Add schema monitoring to your SEO stack. We use a combination of Screaming Frog for crawls (detecting missing datePublished or author fields) and Google Search Console's Rich Results report. But the real test is whether AI search engines are actually citing you. Set up UTM-tagged queries in Perplexity and ChatGPT Search at least weekly. Track which of your pages appear in AI responses. We found that 23% of correctly marked-up content still doesn't get cited if the topic is too niche—AI search has freshness preferences that traditional SEO doesn't.

Schema markup for AI search isn't about rankings anymore—it's about being *findable* in a post-blue-link world. If your schema is missing or incomplete, AI search engines simply won't know what your content is about.

← Back to all articles