AEO for local business: how to get cited when buyers ask AI for recommendations

The buyer asking ChatGPT which plumber to call in their city is one step from dialing. There is no second page to scroll through. There is no comparison map with three options side by side. The AI gives one name, maybe two, and the call goes to whoever appears. The businesses that do not appear never existed in that moment.

This is the new structure of local service discovery. Buyers who used to open Google Maps are now opening ChatGPT or Perplexity first. They ask a plain-language question. An AI model gives a recommendation. If that recommendation does not include your business, you did not lose the comparison. You were never part of it.

The good news: AI models do not choose randomly. They follow a logic you can learn and build toward. This article covers what that logic is, how it applies specifically to local businesses, and what you can do this month to improve your position in AI-generated local recommendations.

Why local AI search changes the stakes

In traditional local search, a Google Maps pack showed three businesses side by side. Buyers compared them. If you made the pack, your odds were roughly one in three.

AI search compresses those odds further. ChatGPT does not default to showing three options. Perplexity might list two or three with context, but the first name carries most of the weight. Google AI Overviews often names a single business or category leader for a local query.

The compression matters because local buyers have high intent. Someone asking Perplexity which HVAC company to call in Charlotte is not browsing options. They are making a decision. The business that gets named wins the appointment. The businesses that do not get named do not exist in that moment for that buyer.

This reality hits small and mid-sized local businesses harder than national brands. A national brand that does not get cited by an AI model loses a fraction of its discovery traffic across many markets. A local plumber that gets passed over loses the call. There is no other market to absorb the miss.

There is also a timing argument for acting now. Most local businesses have not started thinking about AI visibility. National brands have teams building AEO strategies. Your local competitors almost certainly are not doing this work yet. The window to establish your business as the AI-cited option in your market is open right now, and it will narrow as more businesses discover this channel and start building the signals AI models rely on.

How AI models decide which local businesses to recommend

AI models do not have live access to your Google Maps ranking. They do not see your current star rating in real time. What they work from is a combination of retrieved web content and training data, both of which reflect the picture your business presents across the open web.

That picture is assembled from several data sources working together.

Your Google Business Profile provides structured, crawlable data that AI models can parse reliably. Business name, category, service area, hours, and review content are all part of that record. When a model retrieves local data to answer a query, GBP content is among the most consistently structured signals available to it.

Review platforms contribute independently. Yelp, Google Reviews, Angi, HomeAdvisor, the Better Business Bureau, and category-specific directories all publish structured data about local businesses. AI models retrieve from these sources, and the volume and content of your reviews across multiple platforms shapes how confidently a model will recommend you.

Your website schema markup tells AI models what you are, what you do, where you do it, and for whom. A properly implemented LocalBusiness or HVACBusiness or Plumber schema type, combined with Service schema for each distinct offering, gives AI models a structured record they can compare against the intent in a buyer query.

Third-party mentions round out the picture. Local newspaper coverage, industry association listings, regional business directories, and chamber of commerce profiles add independent corroboration. AI models treat these mentions as trust signals. A business that exists only on its own website and its GBP has a thinner signal than one with consistent, corroborated presence across a dozen authoritative sources.

The common thread across all of these inputs is verifiability. AI models recommend businesses they can verify across multiple independent sources. The more consistent and complete your presence, the more confident a model can be in naming you when a buyer asks.

Google Business Profile: your AEO anchor

Your Google Business Profile is the single most important local AEO asset you control directly. It is the canonical record of your business across Google’s ecosystem, which means it feeds Google AI Overviews directly. Third-party AI models also retrieve GBP data when they crawl Google’s public local content.

An optimized GBP does several distinct things for AI visibility.

Business category. Your primary category defines how AI models classify your business against buyer queries. If you run a heating and cooling company and your primary GBP category is “contractor” rather than “HVAC contractor,” AI models may not surface you for heating and cooling queries. Pick the most specific primary category available. Add secondary categories for every distinct service type you offer.

Business description. This 750-character field is your narrative record in AI-readable format. Most businesses fill it with vague marketing language that does not help models characterize what you do. Write for information extraction instead: state clearly what you do, what geographic area you serve, how long you have been in business, and what distinguishes your service. Plain declarative sentences. Specific services named. No claims that cannot be verified.

Q&A section. The Q&A section on your GBP is a FAQ surface that AI models can read and cite. Most businesses leave it empty. That is a missed opportunity. Seed it yourself with the questions your customers ask most often. “Do you offer emergency service?” “What areas do you serve?” “Are you licensed and insured in this state?” “What is your typical response time?” Write each question and answer it accurately and specifically.

Review content. The text of your reviews matters for AI visibility in ways most business owners have not considered. When customers describe the specific service they received, the problem you solved, and the outcome they got, they create extractable content about your business. A review that says “they replaced our water heater in four hours on a Saturday, no extra charge for the weekend call” gives an AI model specific, citable detail. A review that says “great service” gives it nothing to work with. Volume matters, but descriptive depth matters too.

Service menu. The services section of your GBP lets you list every service you provide with a name, description, and price range. This structured service data feeds directly into AI models looking for business-service matching. If you do drain cleaning, water heater installation, leak detection, and pipe repair, each one should be a named service entry with a specific description. A model answering “who does slab leak detection in Austin” is going to find that entry and weigh it.

Profile completeness and recency. GBP profiles with complete photo libraries, verified business attributes, and regularly updated hours signal an actively managed profile. AI models, especially Google AI Overviews, weight completeness and recency in the signals they pull from GBP data. An abandoned profile from 2022 reads differently than one updated this month.

For a detailed look at how GBP optimization connects to local AI visibility, our Google Business Profile service covers the full approach we use with local service clients.

Local schema markup: what to implement first

Schema markup is structured data you add to your website’s HTML that tells AI models precisely what your business is, where it operates, and what it does. For local businesses, four schema types deliver the most immediate AEO impact.

LocalBusiness schema (or its subtypes). Schema.org defines dozens of local business subtypes: Plumber, HVACBusiness, DentalClinic, RoofingContractor, AutoRepair, and many more. Use the most specific type that matches your primary service. This schema should include your business name, full address, phone number, URL, opening hours, and geographic service area. The service area field is particularly important for AI models answering location queries. If your service area is the Phoenix metro, state that explicitly in the schema.

Service schema. For each distinct service you offer, create a separate Service schema block. Each block should name the service clearly, describe it in plain language focused on the outcome it delivers, link it to your business entity as the provider, and specify the service area. A plumbing company might have 10 to 15 separate Service schema blocks, one for each type of plumbing work they handle. This level of detail gives AI models specific signal to match against buyer queries. A model answering “who replaces gas lines in Denver” is going to find a Service schema entry for gas line replacement and weight it heavily.

FAQPage schema. This is the schema type that most directly connects your website content to AI-generated answers. Identify the top 10 to 15 questions your customers ask before hiring you. Answer each one specifically and accurately. Format them as a proper FAQPage schema block in your site’s HTML. When AI models look for answers to local buyer questions, FAQPage schema gives them clean, extractable text they can cite directly. A model answering “do HVAC companies in Raleigh offer financing?” is looking for exactly this type of structured content.

Organization schema with sameAs. The sameAs property connects your business entity to its external presence. List every authoritative profile your business maintains: your GBP URL, your Yelp page, your BBB listing, your Angi profile, your industry association directory page. These connections create a web of corroborated entity signals. AI models follow these connections when assessing how well-established a business entity is.

One critical detail applies to all of these schema types: your Name, Address, and Phone must exactly match your GBP and every directory listing. A suite number missing from your schema address, or “St.” where your GBP says “Street,” creates fragmentation in your entity record. Run a validation pass across your schema after you build it, then do a consistency check against your GBP and your five most prominent directory listings.

Review signals: what AI models actually read

Reviews are content that AI models read, parse, and use to characterize your business. This changes what a useful review program looks like for local businesses trying to improve AI visibility.

Think about what a descriptive review contains. The customer describes the problem they had. They name the service they received. They mention how the team behaved on site. They describe the outcome and whether it resolved the issue. They may mention the location or the technician by name. All of that is structured narrative information an AI model can use to build a profile of your business and match it against buyer queries.

Consider two businesses with the same star rating and the same review count. Business A has reviews that say “great service,” “highly recommend,” and “will use again.” Business B has reviews that say things like “they found the leak under our foundation slab in under an hour, fixed it the same day, and the price was fair for the scope of work.” When an AI model answers “who in Dallas handles slab leak repair,” Business B has evidence. Business A has sentiment. Evidence wins.

This changes how you think about collecting reviews. The goal shifts from volume alone to descriptive depth. Your follow-up message to customers who had good experiences can note that it helps other buyers to know what service they called about and how it went. That nudge, without scripting anyone’s words, produces more useful content.

Respond to every review, positive and negative. Your responses are part of the crawlable record about your business. A specific, professional response to a negative review, one that describes what happened and how you resolved it, creates content about your service recovery process. AI models can read that and factor it into how they characterize you.

Platform diversity matters as well. AI models retrieve from multiple sources. A business with 500 Google reviews and nothing on Yelp, Angi, or the BBB has thinner cross-platform presence than one with 200 reviews spread across four platforms. Build presence where your category’s buyers actually leave reviews.

Location-specific content: pages that AI will actually cite

Your website needs location-specific content pages that AI models can retrieve and reference. This is where most local businesses have the biggest gap, because it requires writing substantive content rather than just building listings and profiles.

When an AI model answers a query like “best roofing company in San Antonio,” it looks for pages that discuss San Antonio roofing specifically, answer questions a San Antonio homeowner would ask, and establish that you have real knowledge of that market. A generic services page that mentions “serving San Antonio” in a footer does not do that work.

A strong location-specific content page covers several things.

The specific services available in that location, named precisely. A roofing company’s San Antonio page should list storm damage repair, metal roofing installation, flat roof replacement, gutter systems, and every other service the team handles in that area. Named specifically, not bundled into vague service categories.

The specific conditions and challenges of that local market. What are the roofing challenges particular to San Antonio’s climate? What weather events drive the most emergency calls? What building materials are common in that area’s housing stock? What local code requirements affect projects there? This information is only credible from a business that actually operates in the area, and AI models weight local specificity accordingly. A page that could describe any city equally well does not carry that weight.

Direct answers to the top questions local buyers ask before hiring. What does a roof replacement typically cost in San Antonio? How long does a standard project take? What permits are required for a full replacement? What warranties cover materials and labor? These answers make the page extractable for AI models answering local pricing and process queries.

Named authorship tied to local expertise. A page that attributes content to a real person with specific local experience carries more entity authority than a page with no human attribution. The author’s name becomes an entity signal connecting the content to a verifiable local expert. If your San Antonio team lead has been doing this work for 12 years in that market, say so.

Avoid templated content. AI models have become good at identifying pages that were written by dropping a city name into a generic template. The tell is content that could describe any market equally well. Real location pages reflect actual knowledge of a specific place. That specificity is both more useful to buyers and more credible to AI models assessing whether to cite you.

NAP consistency: the entity foundation

Your Name, Address, and Phone number are the three data points AI models use to identify your business as a specific, distinct entity. When those three data points match consistently across your website, your GBP, and every directory listing on the open web, you send a clear, strong entity signal. When they do not match, you fragment the picture.

This matters more than most business owners realize. AI models build recommendation confidence through corroboration. When the same business name, address, and phone number appear on your website, GBP, Yelp, Angi, BBB, your industry association directory, and your local chamber of commerce, the model sees multiple independent sources confirming the same entity. That corroboration is what gives a model confidence to name you in an answer.

When your Yelp listing has an old address, your Angi profile uses a different phone number, and your website uses a slightly different business name variant, the model sees contradictory signals. That contradiction works against you when the model decides how confidently to recommend you.

The fix starts with an audit. List every place your business appears online. Check name, address, and phone in each one. Fix every inconsistency. This is methodical work, not complex work, and it is foundational to everything else you build for AI visibility.

Pay careful attention to name variants. “Capitol City Plumbing LLC,” “Capitol City Plumbing,” and “Capitol City Plumbing and Drain” are three different strings representing the same business. AI models need consistency to recognize them as the same entity. Choose one canonical name format and use it everywhere.

Address format also matters. “123 Main St.” and “123 Main Street” are technically identical, but some entity resolution systems treat them as different records. Pick one format and apply it consistently across every listing you maintain.

llms.txt: telling AI models about your business directly

A relatively new but increasingly relevant tactic for local businesses is implementing an llms.txt file on your website.

The llms.txt standard is a plain-text file you place at your domain root that gives AI models a structured overview of your organization, your key pages, and your core services. Think of it as an orientation document for AI crawlers. Rather than making them piece together who you are by parsing your full site, you hand them a summary.

For a local business, an effective llms.txt file includes several things.

A one-paragraph description of your business: what you do, where you operate, how long you have been in business, and what distinguishes your service. Written in plain, declarative language. Specific, not vague.

A list of your most important pages, each paired with a one-line description. Your homepage, your core service pages, your location pages, your FAQ page. Brief descriptions that tell a crawler what each page covers.

Your core service categories named explicitly. If you do HVAC maintenance, HVAC repair, new system installation, and duct work, name all four. If you offer emergency service, say so.

Your service area described clearly. City names, county names, and metro area if you serve a broader region.

Your NAP data. Business name, full address, phone number. Consistent with your GBP.

This file is not a traditional ranking signal for search engines. Its purpose is to help AI models that crawl your site understand your business without having to parse your full HTML. The setup takes less than an hour. The signal it creates is real and increasingly read by AI crawlers that follow this emerging standard.

How to measure AI visibility as a local business

The simplest starting point for measuring AI visibility is manual queries. Ask the AI models your buyers use the questions your buyers ask. You will get direct, immediate feedback on whether you are being cited.

Build a list of 20 to 30 queries your buyers would actually use. Mix them across several formats:

• “Who are the best [service type] companies in [your city]?”
• “What [service type] company should I call in [your city]?”
• “[Service type] near [specific neighborhood]”
• “Who is the most reliable [service type] in [your metro area]?”
• “[Specific problem] in [your city], who do I call?”

Run each query in ChatGPT, Perplexity, and Google AI Overviews. Record what comes back. Note every business cited. Run each query at least three times per platform, because AI models vary their responses between runs. Use the aggregate picture across all runs, not any single answer.

If your business appears in fewer than a third of the responses for queries directly in your service area, you have an AI visibility gap. That gap almost always traces back to one of the issues covered in this article: incomplete GBP data, inconsistent NAP across directories, missing or invalid schema, thin review volume or generic review content, no location-specific content pages, or AI crawlers blocked in your robots.txt.

Repeat the test monthly with the same query set. AI models update their retrieval and training data continuously, and your citation rate will change over time as you build the signals described here. Monthly tracking shows you what is working and where to apply effort next.

Track what AI models say about you when they do cite you, not just whether they cite you. If a model names your business but describes your services inaccurately, or attributes services to you that you stopped offering, that is a content and schema correction to make. The content AI models retrieve to describe your business comes from your own pages and third-party sources, and both can be corrected.

What a professional AEO assessment covers for local businesses

Most of the work described in this article is executable without outside help. A business owner or in-house marketing team that commits to this work can build a strong AI visibility foundation in a few focused months.

Where outside expertise becomes valuable is in the audit and the entity work. A full AEO assessment maps your current citation rate across AI models, identifies specific gaps in your entity signals, audits your schema implementation for errors and omissions, checks whether AI crawlers can access your site, and benchmarks you against the competitors who are already showing up in answers where you are not.

The entity work is harder to execute on your own. Building consistent presence across directories, managing your knowledge graph signals, and structuring your brand’s online footprint for maximum AI recognition requires systematic effort that falls between most marketing teams’ regular responsibilities. It is the kind of work that gets deferred because it is important but not urgent, until the day a competitor starts showing up in every AI answer for your category and you are nowhere.

The schema implementation also has real technical complexity. Common mistakes include mismatched NAP data between schema and GBP, missing sameAs connections, incomplete service descriptions, and FAQPage schema that references questions not actually present on the page. Any of these errors reduce the value of your schema work.

If you want to see exactly where your local business stands with AI visibility, which competitors are being cited in your market, and a specific roadmap for building your citation rate, the AI Visibility and AEO service at AEO Hunt covers the full assessment and implementation plan.

Starting today

The path from invisible to cited follows a clear sequence for most local businesses.

Complete your GBP fully. Every field filled out. Correct primary category selected, and secondary categories added for every service you offer. Business description written for information extraction. Q&A section populated with the questions buyers ask most. A process in place that encourages customers to leave descriptive reviews.

Standardize your NAP. List every directory and listing where your business appears. Fix every inconsistency in name, address, and phone number. This is foundational to everything else.

Add LocalBusiness schema and Service schema to your website. Use the most specific business type subtype that fits you. Create a separate Service schema entry for each distinct service you offer. Add FAQPage schema to every page where you answer common customer questions. Include sameAs links connecting your entity to every authoritative profile you maintain.

Write at least one substantive location-specific content page for your primary service area. Cover the specific services you offer there, the specific conditions of that local market, and direct answers to the top questions buyers ask before hiring. Name the author with their actual credentials and local experience.

Run the AI query test. Search the top 10 to 15 queries your buyers use and see where you stand today. Record the results. That baseline is what you measure future progress against.

Add an llms.txt file to your domain root. One page, covering your business description, your key service pages, your service area, and your NAP. It takes less than an hour to write and the signal it creates is real.

AI models update over weeks and months, not overnight. The results of this work compound gradually. The businesses starting today are building a position that will be very difficult for competitors to unseat once AI models have incorporated these signals and started citing them consistently. Local AI search is early enough that the window is open. But it will not stay that way indefinitely. The time to start is before your competitor does.