AI Marketing • 2026-05-12

Local Business Schema Markup for AI Search: The 2026 Playbook for Showing Up in AI Overviews, ChatGPT, and Perplexity

If you run a local business, the search box stopped being the front door a while ago. The new front door is an AI answer. Someone types "best HVAC company in Fort Myers" or "dentist open Saturday near me," and before they see a single map pin or website link, a block of AI generated text tells them who to pick. That AI is reading something to write that answer. The question every local owner needs to ask in 2026 is simple. Is it reading you, or is it reading your competitor?

Q1 2026 data from Stackmatix shows Google AI Overviews now trigger on 38% of local service queries. Plumbers near me. Roofers in (your city). Med spa open today. Almost four in ten of those searches now start with an AI answer instead of a map pack. And here is the part that should make every owner sit up. The AI local packs surface 32% fewer businesses than the old map pack. Fewer slots, higher bar, more invisible competitors.

The businesses that keep showing up in those AI answers are not the ones with the prettiest websites or the biggest ad budgets. They are the ones whose site code tells the AI exactly what they are, who they serve, where they operate, and what people say about them. That is schema markup. And in 2026, it stopped being optional.

This is the field guide. Read it once, send it to whoever owns your website, and use it as the checklist for getting your local business back on the map (the AI map).

What schema markup actually is, in plain English

Schema markup is structured data that lives in the code of your website. You do not see it as a visitor. The AI does. It is a set of labels that tells Google, ChatGPT, Perplexity, and Gemini, "this part of the page is the business name, this part is the phone number, this part is the address, this part is a customer review, this part is a question and the answer to it."

Without schema, an AI has to guess. It crawls your page, reads paragraphs, and tries to piece together what the business is and what it offers. Sometimes it gets it right. Often it does not. When the AI is unsure, it picks the business it is sure about. That business is the one with clean schema.

With schema, the AI does not have to guess. It reads the structured tags, gets a clear answer in a fraction of a second, and uses your business as the source of its answer. That is how you become the recommendation instead of the runner up.

There are three flavors of schema markup that matter for local businesses in 2026. LocalBusiness schema, which describes the business itself. FAQPage schema, which marks up the questions and answers on your site. And Review or AggregateRating schema, which signals what your customers think. Get those three right and you have done more for your AI visibility than 90% of local businesses have done all year.

Why this matters more in 2026 than it ever has

For most of search history, schema was a "nice to have." Sites without it could still rank if their content was strong and their backlinks were solid. Those days are over. Three things shifted in late 2025 and early 2026 that turned schema from a polish step into the price of admission.

First, Google AI Overviews now consume structured data as a primary input. The AI summary at the top of a Google result is not pulled from a single ranking page. It is assembled from multiple sources, and the sources that "win" inclusion are the ones that present their facts in a format the AI can parse instantly. That format is schema. If your competitor has LocalBusiness schema and you do not, the AI will quote your competitor when someone asks Google about your industry in your city.

Second, ChatGPT and Perplexity have become serious local discovery channels. Research published in early 2026 shows ChatGPT only cites about 15% of the pages it retrieves when answering a question. The other 85% get scanned and discarded. The pages that make the citation cut have clear structure, clear sources, and clear entity signals. Schema is the cleanest possible entity signal you can give a language model. Only 11% of domains get cited by both ChatGPT and Perplexity, which means optimizing for one does not automatically get you the other, but schema markup helps with both.

Third, the local pack itself shrank. Google's AI local packs surface fewer results than the classic three pack ever did. There are simply fewer slots to fight over. When the field narrows, the entry requirements rise. The businesses that make the new shorter list are the ones that look unambiguous to an AI. Everyone else gets sorted into "also serves this area" obscurity.

The cost of being invisible is not theoretical. If you run a service business with a 60 mile radius and you do roughly $400 in revenue per new customer, missing one extra AI answer per day is well over $100,000 in lost annual revenue. That is the math nobody talks about. Schema is the cheapest insurance policy against that loss.

The four schema types every local business needs

You do not need to learn every schema type that exists. There are hundreds, and most of them do not apply to a local service business. You need four. Get these four right and you are ahead of nearly every competitor in your zip code.

1. LocalBusiness schema

This is the foundation. LocalBusiness schema tells search engines and AI systems exactly what your business is. It includes your legal name, the type of business (a more specific subtype like Plumber, Restaurant, Dentist, HomeAndConstructionBusiness), your full address, your phone number, your hours of operation, the geographic area you serve, your website URL, and your social profiles.

The mistake most local owners make is leaving this either missing or half done. A common bad version has the business name and address but no opening hours, no served area, and no business subtype. The AI reads it and thinks "this is a business of some kind, in some category, that may or may not be open." That ambiguity is fatal. The fix is to fill every field. Use the most specific subtype that fits. If you are a roofer, mark yourself as RoofingContractor, not just LocalBusiness. The more specific your label, the more precisely the AI can answer "best roofer near me."

2. FAQPage schema

FAQPage schema marks up question and answer pairs on your site. This is how the old Google "people also ask" boxes used to get populated, and now it is one of the strongest inputs into AI Overviews and ChatGPT responses.

The play here is to identify the questions a prospect actually asks before booking your service. Not generic ones. Specific ones. "Do you charge for estimates?" "How long does a roof replacement take?" "Are you licensed in (your state)?" "What areas do you serve?" Write the questions in the customer's voice, write short clear answers, and wrap them in FAQPage schema. When someone types one of those questions into Google or ChatGPT, your business becomes the source of the answer. That is the inverse of guessing what to write a blog post about. Write the answers your customers are already asking for.

The freshness research from Perplexity is relevant here. Their citation engine prefers content published or updated in the last 30 days at an 82% rate. FAQ blocks are easy to refresh. Update one a month. Date stamp it. That tiny ritual keeps the schema fresh and the citation rate high.

3. Review and AggregateRating schema

If you have collected reviews, you should be marking them up. AggregateRating schema declares your overall star rating and review count in a way that AI systems can read directly. Individual Review schema marks up specific testimonials with the reviewer's name (or a permitted anonymized form) and the rating they gave.

There is a caveat. Google has tightened its policies around what review schema can appear in search results. Self collected reviews displayed only on your own site are fine for AI consumption (LLMs read them) but may not generate rich result stars in Google search. That is okay. The goal is not the star icon. The goal is for the AI answer to say "X is rated 4.9 stars based on 240 reviews" when someone asks about your business or category. Schema makes that sentence possible.

4. Service or Product schema (if it fits)

If your business offers distinct services (water heater installation, mole removal, lash extensions, foundation repair), each major service deserves its own page and Service schema. Service schema lets you describe what the service is, the area it covers, the provider (you), and the type of customers it is for. AI systems love this because it lets them match a query like "emergency water heater repair Cape Coral" to a specific service page from a specific business, not just a homepage.

If you sell physical products, swap Service for Product schema. Same principle.

How AI systems actually read this and pick winners

This is the part most "schema guides" online get wrong. They tell you to install schema and stop there. But schema is one input in a multi step pipeline that an AI runs every time someone asks it a question. Understanding the pipeline helps you understand why certain businesses win.

Step one. The AI retrieves a set of candidate pages from its index. Google does this from the live web. ChatGPT does this from its training data plus its retrieval layer. Perplexity does this fresh on every query. The retrieval layer rewards pages that are crawlable, fast loading, and clearly themed. Schema helps here because it makes the theme obvious.

Step two. The AI parses each retrieved page. It pulls out the structured signals first (schema, headings, lists) and then reads the unstructured body. Pages with schema get parsed faster and more accurately. This means pages with schema get a higher confidence score in the AI's internal ranking.

Step three. The AI generates an answer by stitching together facts from the high confidence sources. Pages with conflicting or missing signals are quietly downweighted. Pages with clean schema get cited more often, both in the visible citation list and in the unattributed paraphrase that often makes up the bulk of the AI's response.

Step four (for some platforms). The AI decides which sources to surface as clickable citations. Only about 15% of retrieved sources get this treatment in ChatGPT. The ones that do tend to be the ones with the cleanest structure and clearest entity claims. Schema is the cheapest way to be in that 15%.

This is why two competitors in the same city can have nearly identical websites and get wildly different AI exposure. The one whose code is parseable wins. The one whose code is mush loses. The visitor never sees the difference. The AI always does.

Common schema mistakes that kill local AI visibility

Schema only works if it is correct. A surprisingly large fraction of local sites have schema installed but installed wrong, and broken schema is sometimes worse than no schema. Here are the mistakes we see over and over in audits.

The first is name, address, and phone number (NAP) inconsistency. Your schema says one thing, your Google Business Profile says another, your Yelp listing says a third. The AI sees the conflict and downweights all three. The fix is to pick one canonical version of your business name, address, and phone number and use it everywhere, byte for byte.

The second is generic LocalBusiness when a specific subtype exists. If you are a dental practice and your schema says LocalBusiness instead of Dentist, you are leaving precision on the table. Match the subtype to the business.

The third is fake or self generated reviews wrapped in Review schema. Google can detect this and will penalize the site. Worse, when an AI parses it and senses the inflation, it learns not to trust your entire domain. Only mark up reviews you actually collected from real customers.

The fourth is FAQ schema with answers that do not match the visible page content. Google explicitly forbids this and has been removing FAQ rich results for offenders since mid 2025. AI systems treat the discrepancy as a trust signal failure. Whatever is in the schema must match what a visitor sees on the page.

The fifth is forgetting to update schema when the business changes. New phone number. New hours. New service area. New owner. Schema does not update itself. If your hours schema still says you close at 5 and you actually close at 7, you have just trained the AI to tell customers you are closed when you are open.

How to install schema if you have a website

You have three realistic options depending on what your site is built on.

Option one. If your site is on WordPress, install a structured data plugin (Rank Math and Yoast both handle LocalBusiness, FAQPage, and Review schema cleanly). Fill in the fields. The plugin writes the JSON LD into your page head automatically.

Option two. If your site is on a hosted builder (Wix, Squarespace, Shopify, GoDaddy Website Builder), the platform may have built in schema tools, but they are often limited. You can supplement by pasting custom JSON LD into a code injection block. This requires generating the JSON, which is easier than it sounds. Tools like Schema.org's own generator, Google's Structured Data Markup Helper, and Merkle's schema generator can produce the JSON for you. Paste it into the page head and you are done.

Option three. If your site is custom built, ask your developer to add JSON LD scripts to the relevant pages. Homepage gets LocalBusiness. Service pages get Service. Pages with FAQ sections get FAQPage. About or testimonials page gets Review and AggregateRating. The developer should know how. If they do not, find someone who does, this is not exotic.

After installation, validate. Google's Rich Results Test (search "Google Rich Results Test") will tell you in seconds whether the schema parses correctly. Schema.org's validator catches edge cases the Google tool misses. Run both. Fix anything flagged. Run again. Done is parseable, not just installed.

What you should be doing this week if your schema is missing or broken

If your local business does not yet have proper schema, the gap is not theoretical. Every week you stay invisible to AI answers is a week your competitors are getting recommended in your place. Here is the practical sequence.

Day one. Pull up your homepage in Google's Rich Results Test. If it returns no detected schema or shows errors, you have your starting point. Screenshot it. This is your baseline.

Day two. Decide on your canonical NAP (name, address, phone). Cross check it against your Google Business Profile, Yelp, Bing Places, Apple Maps Connect, and any directory you appear in. Pick the version you want to be the truth.

Day three. Pick your LocalBusiness subtype. Search "schema.org local business types" and pick the most specific one that fits.

Day four. Write or generate the LocalBusiness JSON LD. Paste it into your homepage head. Validate.

Day five. List the top 10 questions your customers ask before booking. Write the answers. Wrap in FAQPage schema on a relevant page (often the homepage or a dedicated FAQ page). Validate.

Day six. If you have collected reviews (and you should, please ask for them), add AggregateRating schema. If you have a handful of named testimonials with permission, mark them up as Review schema.

Day seven. Update your Google Business Profile to match. Post a photo. Reply to a review. The freshness signal compounds the schema signal. Profiles that go 30 plus days without activity now drop in rankings, and the algorithm reward for "interactions" is real. If you want a deeper read on what is driving local rankings right now, see how local businesses show up in ChatGPT and AI search in 2026 and the breakdown of Google's upgraded AI Overviews and the 48% query share they now command.

That is one week. One week to go from invisible to AI ready. Most local businesses will never do it. The ones that do will quietly own their category.

How to know if it is working

Schema is not a "set it and watch the chart" play. Most of the impact is invisible to traditional analytics because the AI answer never sends a click. Here is how to measure real lift.

Track AI Overview impressions in Google Search Console. The AI Overview clicks and impressions show up as a distinct surface in the latest GSC reports. If your impressions on that surface climb in the 30 days after schema goes live, that is signal.

Track ChatGPT and Perplexity referrals in your analytics. Both now send identifiable referrer traffic. Watch for chatgpt.com and perplexity.ai in your top sources. Even a small trickle is meaningful, those clicks tend to convert better than cold ad traffic because the AI already vouched for you.

Track direct branded search. When your business gets cited by AI more often, your branded search volume rises. People hear about you in an AI answer, do not click, and search your name later. That is the lagging signal that proves the AI is doing the work.

Track Google Business Profile call clicks and direction requests. Strong AI presence usually drives more profile interactions, not fewer. If those numbers climb after schema goes live, the AI loop is closing.

Track conversion quality. AI sourced visitors come prequalified. They have already been told you are the right answer. Watch your booking or lead form conversion rate from those sources. Higher than your site average is normal.

FAQ

Do I need schema if I already have a strong Google Business Profile?

Yes. Your Business Profile lives on Google's surfaces. Schema lives on your website and is what ChatGPT, Perplexity, Gemini, and Google's AI Overviews read off your site directly. The two work together. Doing only one is leaving half the AI ecosystem out of reach.

Will schema help my regular SEO too?

Yes. Schema has been a recognized ranking input on Google for years, and the AI shift only amplifies it. Sites with strong schema typically pick up better rich result eligibility, higher click through rates, and clearer category targeting in traditional search alongside the AI gains.

Can I install schema myself if I am not technical?

Yes if your site is on a major platform with a plugin or built in schema support. WordPress with Rank Math or Yoast is the easiest case. If your site is custom or on a niche builder, you will likely need help, this is the right time to call in a specialist.

How long until I see results?

Google typically begins surfacing new schema within days. ChatGPT and Perplexity update on their own retraining or crawl cycles, which can range from a few weeks to a few months. Realistic expectation is meaningful lift within 30 to 90 days, with compounding gains the longer the schema stays clean and current.

Is schema markup against any AI platform's rules?

No. Schema is an open web standard that Google, Microsoft, Yandex, and now every major AI search system explicitly support. Using it correctly is encouraged. Using it deceptively (fake reviews, mismatched FAQs) is what gets penalized.

What happens if my schema breaks after a site change?

You quietly lose AI visibility, often without realizing it. This is why a quarterly schema audit matters. Most local businesses change a phone number, swap a CMS theme, or migrate a page and never re validate. That is how AI presence silently erodes.

Will my competitor copy my schema?

They might try. Schema is public, anyone can view your page source. The advantage is not the schema itself, it is the underlying business clarity (clear services, real reviews, fresh content) that the schema reflects. Copying the tags without the substance does not work, the AI cross checks.

The bigger picture

Local search used to reward two things. A polished Google Business Profile, and a website that did not embarrass you. That world is gone. Local search in 2026 rewards businesses that have made themselves legible to machines. Machines that summarize, recommend, and rank without ever showing a human the source.

Schema markup is the language those machines read. The local businesses that learn it early will own their categories for years before everyone else catches up. The ones that wait will keep wondering why their phone is quieter than it used to be, even though the website "looks great."

The 38% AI Overview share for local services is not a peak. It is a floor. Every quarter, more queries move into AI answers. Every quarter, the bar for inclusion rises. The fix is straightforward, and you do not have to do it alone.

We help local businesses get found in AI search by tightening the structured data, schema, and content signals that AI systems actually read. If your business feels invisible online lately, that is usually what is happening under the hood. → aibizit.com