Schema Markup: Getting Your Business Found in AI Search Results

We've been testing schema markup performance across traditional Google Search, ChatGPT Search, and Perplexity for the past 6 months. Here's what's changed: schema markup has gone from "nice to have" to essential. AI search engines rely on structured data to pull information directly from your site without having to parse messy HTML. A dentist we work with saw their practice appear in ChatGPT Search results for 23 different queries after implementing LocalBusiness + MedicalBusiness schema. They weren't ranking on Google for 15 of those queries. That's new traffic AI search isn't cannibalizing from Google—it's incremental.

The Schema Types AI Engines Actually Use

Not all schema markup is equal in the age of AI search. We've monitored which data types appear in AI-generated answers and found that AI engines prioritize LocalBusiness, Organization, Product, Review, and Service schema. For local service businesses, LocalBusiness + Service schema combination is performing best. A plumbing company we track added Service schema markup for "emergency plumbing," "leak repair," and "water heater installation"—within 30 days, they appeared in Perplexity answers for 8 different service-specific queries.

• LocalBusiness: name, address, phone, hours, service area (most important for local)
• Service: service name, description, provider, areaServed, priceRange
• Review + AggregateRating: cite count matters (50+ reviews beats 5)
• FAQPage + Question: directly answers common queries AI engines pull from
• Product (for retail): price, availability, review, image

FAQPage Schema: The Fastest Way to AI Visibility

We tested FAQPage schema across 15 websites and found it was the most reliable way to get extracted into AI search results. ChatGPT Search, in particular, pulls from FAQ schema to build answers. A tax prep firm we work with implemented 22 FAQ items covering "self-employed tax deductions," "quarterly estimated taxes," and "home office deductions." Within 8 weeks, they showed up in ChatGPT answers for 19 of those 22 topics. The traffic didn't spike massively—maybe 40-60 clicks per month from ChatGPT—but it was qualified traffic and positioned them as an authority.

"FAQ schema is like handing AI search engines a pre-written answer about your business. They don't have to guess or infer—they have the exact information, formatted correctly."

The trick is writing FAQ items that match how people actually search in AI engines. Don't write "Why should you choose us?" Write "What's the difference between a standard roof inspection and a comprehensive inspection?" or "How long does HVAC maintenance take?" These are questions real customers ask. Use your support tickets and customer emails as inspiration. If customers ask it 5+ times, it's FAQ material.

Implementation: Where Most Businesses Go Wrong

We've audited 40+ sites in the past quarter and found that 65% of them have broken or incomplete schema markup. The most common mistakes: schema is implemented on one page and nowhere else (should be site-wide where relevant), LocalBusiness schema lists the wrong address or phone, or Review schema pulls from reviews older than 2 years. AI engines are smarter about staleness now—old data gets lower weight.

• Use Google's Rich Results Test to validate schema before publishing
• Implement schema on key pages: homepage (Organization + LocalBusiness), service pages (Service), testimonial/review pages (Review + AggregateRating)
• Keep Review dates current—remove old reviews from schema after 18 months
• Test across Bing, ChatGPT, and Perplexity (not just Google) to see extraction
• Update schema quarterly to reflect current pricing, hours, and service offerings

The Competitive Moat: Who's Winning

Right now, businesses with complete, accurate schema have a genuine advantage in AI search. We compared 8 competing plumbing companies in one market—only 2 had proper LocalBusiness + Service schema. Those 2 dominated AI search results for service queries, even though they weren't the highest-ranking on Google Search itself. This won't last forever (schema will become table stakes), but for the next 12-18 months, it's a competitive gap worth exploiting.

← Back to all articles