Most local business owners we talk to in Huntington Beach already have a website that ranks for at least a few of their core terms. What they don’t have is schema for ai search — the structured data that tells ChatGPT, Perplexity, Google AI Overviews, Gemini, and Claude what your business actually is, who it serves, and why it deserves a citation in the answer box.
That gap is the biggest visibility lever a local service business has in 2026. Schema is no longer a Google ranking nudge — it’s the machine-readable resume your business shows to every AI model deciding whether to mention you by name. This guide is the exact setup we ship for dentists, contractors, HVAC companies, and restaurants across Orange County. Plain code, no fluff, no $5,000 audit upsell.
Why Schema for AI Search Matters More Than Schema for Google in 2026
Schema markup has been around since 2011. For most of its existence the payoff was modest: a rich snippet here, a star rating there, a faster path to the featured snippet. Google still uses it that way, but the rules changed when ChatGPT, Perplexity, and Google AI Overviews started generating answers instead of returning links.
AI models don’t read your homepage the way a human does. They parse the structured data first, then check the surrounding content for confirmation. Without schema for ai search, your business is invisible to that parsing layer. With it, you become a named entity that AI models can quote, cite, and recommend.
The other shift is volume. Google’s own data shows AI Overviews already appearing on a significant share of informational and comparison queries. For a local service business, “best HVAC company in Costa Mesa” or “how much does Invisalign cost in Huntington Beach” is now an AI-answered query first, a ten-blue-link query second.
What AI Overviews Actually Read on Your Page
Google’s AI Overviews don’t just summarize the top 10 organic results. They cross-reference structured data, Google Business Profile signals, and entity graphs from Wikidata and other knowledge bases. If your structured data is missing, contradictory, or generic, the model deprioritizes you in favor of a competitor whose data is clean.
Three signals matter most for AI Overview inclusion:
- Entity identity — does your
LocalBusinessschema match your Google Business Profile name, address, phone, and category exactly? - Service specificity — do you mark up the actual services you offer (drain cleaning, root canal, HVAC repair) rather than a vague “professional services” label?
- Answer-ready content — does your
FAQPageschema use the same questions real people search for, in conversational language?
The same checks apply to Perplexity and ChatGPT, which both pull from structured data when answering local-intent queries. This is why we treat schema as the foundation of any generative engine optimization project, not as a technical SEO afterthought.
The Schema for AI Search Starter Pack: Four Types Every Local Business Needs
Most local service business websites we audit in Orange County are missing at least three of these. The fix usually takes one focused afternoon.
1. LocalBusiness schema
This is the entity ID for your business. Without it, AI models have no canonical record of who you are. Include name, address, phone, geo-coordinates, hours, primary category, and the URLs of your social and review profiles. Use the same NAP that appears on your Google Business Profile — letter for letter.
2. Service schema
One block per service you offer. A dental practice should have separate Service blocks for cleanings, fillings, implants, and Invisalign. An HVAC company needs blocks for installation, repair, and maintenance. Generic “we do everything” pages get ignored by AI models because they can’t pull a specific answer out of them.
3. FAQPage schema
This is the highest-ROI schema type for AI Overviews in 2026. Each FAQ block should mirror the exact phrasing of a real search query. AI models pull these as direct quotes when generating answers. Keep each answer between 40 and 80 words.
4. Review and AggregateRating schema
If you have a 5-star Google profile with 50+ reviews, mark it up. AI models use review signals as a trust threshold before citing a business by name. The schema doesn’t fabricate reviews — it exposes the real ones to the parsing layer so they actually count.
The Four Schema Types Most Local Businesses Are Missing
Beyond the starter pack, four types meaningfully move AI Overview citation rates for local service businesses. These are the ones we add during local SEO services for service businesses onboarding:
- Person schema for the founder or lead practitioner. Names tied to credentials (DDS, Ph.D., licensed contractor) signal authority to AI models. This is the same passage-level trust signal that makes brand mentions replace backlinks for AI search.
- BreadcrumbList schema on every service and location page. Lets AI models map your URL hierarchy without guessing.
- GeoCoordinates schema for service-area businesses without a storefront. Tells AI models the radius you actually serve.
- HowTo schema for any guide-style content. Models love structured steps because they translate directly into AI Overview bullet lists.
We see roughly a 20-40% lift in AI citation impressions within 60 days after these four are installed cleanly. Not a guaranteed number, but consistent enough across the OC dentist, contractor, and HVAC clients we track that we now consider it the floor of a competent install.
How to Install Schema Without a Developer
You don’t need a JavaScript build pipeline. Pick one of three paths based on your platform:
- WordPress. Install RankMath or Yoast SEO, enable the schema module, fill in the LocalBusiness fields. Add FAQs through the block editor. Done.
- Squarespace, Wix, or Shopify. Use a code injection block in the site header. Paste a hand-rolled JSON-LD block. Validate before publishing.
- Custom site (Astro, Next, Webflow). Add JSON-LD directly to your layout component. We do this for every client website we build because it survives template changes and ships with every new page automatically.
We’ve documented the exact JSON-LD blocks we ship for OC clients inside our Google Business Profile optimization checklist — the schema work pairs with the GBP cleanup because the two have to match exactly, or AI models treat the mismatch as a signal the business can’t be verified.
How to Test Your Schema After You Ship It
Always validate. Two tools, both free:
- Google’s Rich Results Test confirms Google can parse it.
- The official Schema.org validator confirms the syntax is correct against the spec.
Run both before pushing to production. Then, two weeks later, pull a Search Console performance report filtered to AI Overview impressions. If those impressions haven’t moved, the schema is parsing but the surrounding content needs a passage-clarity pass. That’s a different problem with a different fix.
Common Schema Mistakes That Cancel Out AI Overview Citations
Most schema we audit fails for one of five reasons, and every one is fixable in under an hour:
- NAP drift — the address in schema doesn’t match the Google Business Profile. Instant disqualification.
- Fake reviews — invented
AggregateRatingblocks with no real source. Google penalizes this and AI models stop trusting the domain. - Stale hours —
openingHoursfrom 2022 still in the markup. Update quarterly. - Missing
@id— without a stable@id, AI models can’t link yourLocalBusinessto yourServiceandPersonblocks. Each entity needs a URL-based@id. - Schema buried in JS — if your structured data only renders client-side, half the AI crawlers can’t see it. Server-render JSON-LD.
What “Good” Looks Like 60 Days After Install
A clean schema rollout on a local service business site should show, by day 60:
- AI Overview impressions trending up in Search Console for your business name and 2-3 service+city queries
- ChatGPT and Perplexity citing your business by name when a user asks for OC service comparisons
- Zero structured data errors flagged in Search Console
- LocalBusiness schema matching Google Business Profile letter for letter
- All four core schema types — LocalBusiness, Service, FAQPage, Review — shipped and validated
If you’re 60 days past install and none of this has happened, the schema is parsing but isn’t connected to an entity AI models trust. Usually a brand-mention or topical-authority problem — not a code problem. That’s where the SEO + GEO package work picks up.
FAQs About Schema for AI Search
Is schema for AI search the same thing as GEO? No, but they overlap. GEO is the broader discipline of optimizing for inclusion in AI-generated answers. Schema is one of the core inputs GEO uses — alongside passage-level content structuring, brand mentions, and crawler accessibility.
Does schema help with ChatGPT and Perplexity, or only Google? Both. ChatGPT, Perplexity, Claude, and Gemini all parse structured data either directly through their crawlers or indirectly through the indexes they license. Clean schema lifts citation rates across the board.
How much does it cost to have an agency install schema? For a single-location local business, expect $400-$900 for a one-time install, or it’s bundled into a monthly local SEO retainer. Anything above $2,500 for schema alone is overpriced unless the site has 50+ service or location pages.
Do I need to redo my schema every year? Audit once a quarter. Update when hours, services, or NAP change. A full rewrite is rarely needed unless the underlying website is being rebuilt.
What about voice search and Apple Maps? The same structured data feeds Siri, Alexa, and Apple’s local listings. Schema is platform-agnostic — install it once, every assistant that pulls from your domain benefits.
Photos by Zulfugar Karimov, Marvin Meyer, Luke Chesser and Bernd 📷 Dittrich on Unsplash.
