AEO for Local Business: Get Cited by AI Near You

AEO for local business is answer engine optimization applied to a place: your shop, clinic, salon, restaurant or service area. The goal is simple to state. When someone near you asks an AI tool which business to choose, your name should be in the answer, described correctly, with a link back to you.

To keep this concrete, we’ll follow one example through the whole guide. Meet Paws & Bubbles, a made-up pet grooming salon in Bangkok’s Sathorn district. It has a Google Business Profile with decent reviews, a basic website and a LINE Official Account where most bookings happen. Business is fine. But the owner has noticed that some new customers say “ChatGPT told me about you,” and others mention competitors they found the same way.

Now picture a pet owner who types into ChatGPT: “Good dog groomer in Sathorn for an anxious senior poodle, open Sunday, with clear prices?” ChatGPT doesn’t show ten links. It looks at where the person is, searches the web, reads some listings, reviews and websites, and writes back three or four names with a short reason for each.

AEO for a local business is everything Paws & Bubbles does to be one of those names, and to be described accurately when it is. That covers its Google Business Profile, its listings on other sites, its reviews, its own web pages and what other websites say about it.

If the term AEO is new to you, our plain-English guide to what answer engine optimization is covers the basics. And if you want to know how to track whether any of this is working, our pillar on measuring AI visibility and tracking AI citations is the companion to this guide.

Because an AI answer has far fewer slots than a map, and the AI is much fussier about who fills them.

The clearest data comes from SOCi’s 2026 Local Visibility Index, released on 28 January 2026. SOCi analyzed more than 350,000 locations across 2,751 multi-location brands and compared how often they appeared in Google’s local 3-pack with how often AI assistants recommended them (SOCi).

The gap is large. Brands appeared in Google’s local 3-pack 35.9% of the time on average. ChatGPT recommended just 1.2% of locations. Gemini recommended 11% and Perplexity 7.4%. SOCi estimates AI visibility is three to 30 times harder to achieve than ranking well in traditional local search (Search Engine Land).

Two other findings from SOCi are worth knowing. Only 45% of the retail brands that led traditional local search were also among the brands AI recommended most, so map strength alone doesn’t carry over. And the locations ChatGPT did recommend averaged 4.3-star ratings, which tells you reputation is part of the filter.

SOCi’s sample is big chains, not independent shops. A newer study helps fill that in. In September 2026, BrightLocal published a Local AI Visibility Study based on about 200,000 localized, non-branded searches for roughly 1,300 business locations, producing about 1.9 million AI citations. Google Maps mentioned a tracked business in 66% of searches. The AI surfaces mentioned it far less often: 38% in AI Overviews, 33% in ChatGPT and 32% in AI Mode (BrightLocal).

Why the difference? A map shows many pins and lets you scroll. An AI answer is a short piece of writing. BrightLocal found AI responses named about 4.1 businesses on average in ChatGPT, 3.5 in AI Mode and 2.5 in AI Overviews. Fewer slots means a higher bar.

The answers also move around. On repeat searches, BrightLocal found only 20% to 33% overlap between result sets (23% for ChatGPT, 21% for AI Mode, 33% for AI Overviews), compared with 99% agreement across attempts on Google Maps. So “we’re in ChatGPT” is never a yes-or-no fact. It’s a rate, and the job is to push that rate up.

ChatGPT builds a local answer in a few steps: it estimates where you are, decides to search, reads a set of sources, and writes an answer that names the businesses those sources support. OpenAI doesn’t publish exactly how it ranks them, but the inputs are documented well enough to act on.

Where you are

By default, ChatGPT collects general location information based on your IP address, according to OpenAI’s help center (OpenAI). Since late March 2026, users can also choose to share their device location. The setting is off by default, lives under Settings, Data Controls, and has two layers: approximate and precise. With precise location on, ChatGPT can use your device’s specific location, such as an exact address (PPC Land).

So the same “groomer near me” question can return different names for someone in Sathorn and someone 5 kilometers away in Ari, and different names again when location sharing is off. Keep that in mind when you test.

Where it searches

OpenAI’s help center says ChatGPT search may share disassociated search queries with third-party search providers such as Bing. Its own crawler, OAI-SearchBot, is used “to surface websites in search results in ChatGPT’s search features,” and OpenAI recommends allowing it in robots.txt (OpenAI crawler docs).

Damian Rollison of SOCi compared ChatGPT’s local answers with Bing side by side for Search Engine Land. His tests, plus ChatGPT’s own description of its process, suggested it scans the top 20 to 30 web results, looks more closely at 5 to 8 promising sources, and draws its answer from 3 to 5 of them (Search Engine Land). He stresses this is observation, not an official spec, so treat the numbers as a rough picture.

Which data it licenses

Local answers also draw on licensed data. In its full-year 2025 results on 12 February 2026, Yelp announced an agreement with OpenAI (Yelp). Reporting in July 2026 described the deal as licensing Yelp’s reviews, ratings, photos and business information so ChatGPT can show them inside local answers (PinMeTo). ChatGPT has also shown local results with a map and list view since its December 2024 search update (GSQI).

Search is also open to everyone. In February 2025, OpenAI made ChatGPT search available to all users, including people who aren’t logged in.

What does this mean for Paws & Bubbles? Three things. Its website must be readable by OAI-SearchBot and indexed in Bing, not just Google. Its listings outside Google, Yelp included where Yelp is used, need to be accurate. And the pages ChatGPT is likely to read, such as local best-of lists, need to mention it. For the platform-specific playbook, see our guide on getting cited in ChatGPT answers.

Gemini and Perplexity read the same kinds of sources as ChatGPT, but they weight them differently, so it helps to know each one’s habits.

Gemini

Gemini has a home advantage: it’s grounded in Google Maps. In SOCi’s 2026 study, the profile information Gemini showed was accurate 100% of the time, against about 68% on ChatGPT and Perplexity. Gemini also recommended the most locations of the three AI tools, at 11% (Search Engine Land).

Gemini also likes business websites. Yext analyzed 6.8 million citations from about 1.6 million responses across Gemini, OpenAI and Perplexity models, collected between July and August 2025 for more than 200,000 locations. Gemini drew 52.15% of its citations from first-party websites, the highest of the three. Yext also notes that Gemini uses Google Business Profile data in its answers without citing it, because that data is Google’s own (Yext).

A smaller 2026 study points the same way. Steady Demand, a US local SEO agency, logged 14,472 AI citations from 1,487 queries across 50 large US metros and 10 local service categories. Business websites made up 59.9% of Gemini’s citations. The same study found Gemini named the same top business on only about 7.9% of repeat queries, while Google’s local pack returned the same top listing 90.2% of the time (PinMeTo). It’s an agency study with a modest sample, so use it as a direction, not a law.

Perplexity

Perplexity leans more on directories and review platforms. In Yext’s data, Perplexity sent more than 364,000 listing citations to MapQuest and more than 239,000 review citations to Tripadvisor, and industry-specific directories reached 24% of its citations for unbranded subjective questions like “best groomer near me” (Yext). Perplexity has also used Yelp data for local answers, including maps, photos and review quotes, since March 2024 (Maginative).

For Paws & Bubbles, that means two different priorities. For Gemini, the Google Business Profile and the salon’s own site do most of the work. For Perplexity, consistent profiles on the directories and review sites that cover pet services in its city matter more.

Mostly sources you own or can manage: your website and your listings. That’s good news, because it means a local business has real control.

BrightLocal’s September 2026 study is the largest recent look at this. Across about 1.9 million citations in ChatGPT, AI Mode and AI Overviews, business websites accounted for 42% of all citations and 93% of the unique domains cited. Google Business Profile accounted for about 28.5% of citations. Yelp followed at 9.53%, Facebook at 2.23% and Tripadvisor at 1.79% (BrightLocal).

Yext’s 2025 study reached a similar conclusion from a different angle. Of its 6.8 million citations, more than 2.9 million came from brand websites and more than 2.9 million from listings, while reviews and social added more than 545,000. Yext estimates that brands can directly influence or manage the sources behind about 86% of AI citations. Forums were small, at about 2% (Yext). Note that the data comes from Yext’s clients and prospects, mostly multi-location brands.

The mix shifts by industry. In Yext’s data, listings made up 52.61% of citations in healthcare, while websites led in financial services (48.16%) and retail (47.62%). Food service had the highest review share, at 13.28%. So a restaurant should care a lot about review platforms, while a clinic should care a lot about the health directories that list it.

Older data shows how fast this moves. When BrightLocal ran 800 local searches in ChatGPT in November 2024, business websites made up 58% of sources, and Yelp, Facebook and Google Maps didn’t appear as directory sources at all (BrightLocal). By 2026, Yelp was the third biggest source in its larger study. Platforms change their data deals, so the safe strategy is to be accurate everywhere rather than to chase one source.

Forums are the wild card. Yext measured them at about 2%, but Steady Demand’s 2026 study found Reddit at 13.7% of citations, more than all local directories combined (PinMeTo). The difference likely comes from the categories and query types each study used. Either way, what real people say about you in public threads can end up in an answer.

You get cited by AI as a local business by giving it the same things a careful human researcher would want: a site it can read, a complete and accurate profile, matching details everywhere, recent honest reviews, clear pages about each service and independent sources that vouch for you.

We break the work into six parts. Each gets its own section below, but it helps to see the whole plan first, because the parts depend on each other. Great service pages don’t help much if AI crawlers are blocked. Lots of reviews don’t help if your name and address differ across five directories.

Crawler access. AI tools can only cite what they can fetch. For ChatGPT, that means allowing OAI-SearchBot in robots.txt; OpenAI says changes take about 24 hours to apply for search (OpenAI). Also check that your firewall or CDN isn’t quietly blocking AI crawlers, and that your key pages are indexed in both Google and Bing. Our free AI crawler checker tests this in a minute.

Google Business Profile. This feeds Google’s AI features and Ask Maps directly, and it was the second largest source of local AI citations in BrightLocal’s 2026 study.

Consistent listings. AI tools cross-check. If your hours say one thing on Google and another on Facebook, you give them a reason to doubt you, or to describe you wrongly.

Reviews. Ratings act as a filter, and review text tells the AI what you’re good at.

Service pages. A dedicated page for each service was the second strongest AI factor in Whitespark’s survey.

Local mentions. Best-of lists, local news and association pages were the strongest AI factors of all.

Paws & Bubbles would tackle them roughly in that order, because the early items are fast and cheap and the later items take months. The rest of this guide walks through each one, starting with how to find out where you stand today.

You check by asking the AI tools the questions your customers ask, several times, from where your customers are, and writing down what happens. One screenshot proves little, because local answers change from run to run and from street to street.

Start with the questions. Pull them from real sources: what people ask on the phone, in LINE or WhatsApp chats, in reviews and at the counter. For Paws & Bubbles, that might include:

• “Best dog groomer in Sathorn”
• “Groomer near me for an anxious senior dog”
• “Cat grooming Bangkok open on Sunday”
• “How much is a full groom for a poodle in Bangkok?”
• “Dog groomer near Chong Nonsi with English-speaking staff”

Mix “best X near me” questions with questions about problems, prices and specific needs. AI answers for the specific ones are often less crowded, and that’s where a smaller business can win first.

Then run each question in ChatGPT, Gemini, Google AI Mode and Perplexity. Run each at least three times, because BrightLocal found only 20% to 33% overlap between result sets on repeat searches (BrightLocal). Location matters just as much. BrightLocal found AI answers for points in the same town overlapped only 36.2% in ChatGPT and about 47% in AI Mode and AI Overviews. If you serve several neighborhoods, test from several.

For every answer, record four things: whether you’re named, whether the description is right (hours, prices, services, area), which competitors are named, and which sources are cited. That last column is the most useful. It’s your shopping list. If ChatGPT keeps citing a local “best groomers in Bangkok” article that doesn’t mention you, that article is a target.

This is slow by hand. It’s the reason we built GrixTrack, our tracker at Grixhub, which runs each prompt three times in ChatGPT, Gemini and Google AI Mode and records mentions and citations, alongside Google rankings by country, city and language. It has a free plan with 10 keywords, 5 prompts, 1 project and 2 competitors, which is enough for a small business to get a baseline. You can see it on the GrixTrack page, and our guide on AI visibility and AI share of voice explains the metrics.

They matter because AI tools read many sources about the same business and look for agreement. Listings also make up a huge share of what AI cites: in Yext’s study, more than 2.9 million of 6.8 million citations came from listings, and for OpenAI models listings were 48.73% of all citations (Yext).

Think about it from the AI’s side. It finds Paws & Bubbles on Google with Sunday hours, on Facebook with “closed Sunday,” and on an old directory with the phone number from before the move. Which is right? It can’t call to check. It can either leave the salon out, or pick one and risk telling a customer the wrong thing. Neither helps.

So the job is a cleanup, then a habit.

Make a master record. Write down your exact business name, address format, phone, website, hours, categories and a two-sentence description. Every listing should match it.

Claim the listings that matter. Google Business Profile first. Then the places AI tools are known to cite: Yelp (licensed to OpenAI, and the third biggest source in BrightLocal’s 2026 study), Facebook, Tripadvisor where you’re in hospitality, Apple Maps through Apple Business Connect, and Bing Places, since Bing is one of ChatGPT’s search providers. BrightLocal’s 2026 survey found the share of consumers reading reviews on Apple Maps rose from 14% to 27% in a year (BrightLocal).

Add your industry’s directories. Yext found healthcare answers drew 52.61% of their citations from listings, and Perplexity’s industry-specific directory share reached 24% for some query types. Every industry has its own: health directories for clinics, legal directories for law firms, booking platforms for salons and restaurants.

Fix the old ones. Search your old phone number and old address. Old listings and data aggregator copies often outlive a move by years.

The Thailand version

In Thailand, the list looks a little different. Google Maps is still the base layer. Wongnai is a major listing and review platform for restaurants and beauty businesses. LINE is where many customers contact businesses: LINE Thailand reports 56 million users and more than 12 billion conversations through LINE Official Accounts in 2025 (Relevant Audience, citing LINE Thailand). AI tools don’t read private LINE chats, but your LINE OA ID on your profile and website gives customers the contact route they expect.

There’s also a bilingual wrinkle. Many Thai businesses have a Thai name and an English name, and listings use them inconsistently. Pick how you present both, for example “Paws & Bubbles (พอว์ส แอนด์ บับเบิ้ลส์)”, and use it the same way everywhere, including the address format. Our guide to doing SEO in Thailand covers bilingual search in more depth.

Reviews affect AI recommendations in two ways. They’re a filter, because AI tools tend to recommend well-rated businesses. And they’re a description, because the AI reads review text to learn what you’re good at and repeats those themes in its answer.

The filter is clear in the data. In SOCi’s 2026 study, locations recommended by ChatGPT averaged 4.3 stars, Perplexity’s 4.1 and Gemini’s 3.9. Underperforming financial brands had average ratings near 3.4 stars and replied to fewer than 5% of reviews (Search Engine Land). Whitespark’s experts ranked high Google ratings and the authority of third-party review sites among the top ten AI visibility factors.

The description part is easy to miss but just as important. Google’s own Maps team describes Ask Maps sifting through community reviews to answer specific needs, like finding a quiet bar.

Now apply that to Paws & Bubbles. If twenty reviews mention “patient with my nervous old dog,” an AI has evidence for the “anxious senior dog” question. If reviews only say “great service,” it has nothing specific to repeat. You can’t write your customers’ reviews, and you shouldn’t try, but you can ask in a way that invites detail: “If you have a minute, tell other owners what your dog came in for and how it went.”

What customers expect has also shifted. BrightLocal’s 2026 survey of US consumers gives the clearest benchmarks.

Recency stands out. Because AI review summaries and AI answers lean on what’s current, a business with 300 reviews from three years ago can look weaker than one with 60 reviews from the last six months. Steady beats bursty.

Stay inside the rules

Google’s policy bans reviews “that have been paid for, directly or in kind,” offering incentives such as discounts or free services for reviews, and selectively asking only happy customers for reviews (review gating). It also covers conflicts of interest such as employees reviewing their employer (Google Maps policy). Breaking these rules risks removed reviews and profile penalties, and it fills your profile with exactly the bland, suspicious text that doesn’t help an AI.

The honest system works fine: ask every customer, make it easy with a short link or QR code at the counter and in your LINE follow-up message, and reply to every review, good or bad. BrightLocal found 50% of consumers are put off by templated replies, so write them like a person.

The most useful schema for a local business is LocalBusiness structured data, using the most specific subtype that fits. It isn’t required for AI features, and it won’t make an AI recommend you on its own, but it removes guesswork about your core facts.

Google is clear on both points. For AI Overviews and AI Mode, “there’s also no special schema.org structured data that you need to add,” and any structured data should match the visible text on the page (Google Search Central). For local business markup, Google says you can use it to tell Google about business hours, departments and more. The required properties are address and name, and the recommended ones include geo coordinates, openingHoursSpecification, priceRange, telephone, url, menu and servesCuisine for restaurants. Google also advises using “the most specific LocalBusiness sub-type possible” (Google Search Central).

For Paws & Bubbles, a sensible setup would be:

On the home page or contact page: a LocalBusiness block with the exact name from the master record, the full address, telephone, url, geo coordinates, openingHoursSpecification including any seasonal hours with validFrom and validThrough, and priceRange.

On each service page: a Service block that names the service, links it to the business as the provider and states the area served. Keep the wording identical to the visible page.

On the FAQ section: FAQPage markup only where the questions and answers are visible on the page, word for word.

One rule trips people up. Google’s review snippet guidelines say that if a business controls the reviews about itself, its pages using LocalBusiness or Organization markup “are ineligible for star review feature” (Google Search Central). So marking up testimonials you collect yourself won’t earn stars. Your real reputation lives on Google and third-party review sites, which is where AI looks anyway.

Our free schema generator builds valid LocalBusiness JSON-LD in your browser. Test it with Google’s Rich Results Test before publishing, and update it whenever your hours or address change. Stale schema that contradicts your profile is worse than none.

You earn them the slow, honest way: by being worth recommending, then making sure the people who write local lists, articles and forum answers know about you. This is the part of local AEO that takes longest, and it’s also the part Whitespark’s experts ranked highest. Presence on expert-curated best-of lists was the number one AI visibility factor, and three of the top five factors were about mentions on other sites (Whitespark).

Start with your prompt audit. The sources column shows exactly which lists and articles the AI tools already trust for your category and city. For Paws & Bubbles, that might be a city magazine’s “best groomers” roundup, a pet blog’s neighborhood guide and a few forum threads. Those are your targets.

Best-of lists. Many are written by editors who visit, call or read reviews. Make it easy for them: a short, factual media page with your story, photos, prices and contact details. Avoid paid “award” schemes that sell badges to anyone; they’re rarely the lists AI cites, and some readers see through them.

Local press and blogs. Local journalists want local stories. A salon could share practical advice for pet owners during the rainy season, or host a free nail-trim day for a shelter. Real stories earn real coverage, and each article is an unstructured citation of the kind Whitespark’s experts value.

Associations and community. Trade associations, chambers of commerce, business districts and charity partners often list members on their websites. These are trusted, relevant domains, and the third most important AI factor in Whitespark’s list was prominence on industry-relevant domains.

Forums. Reddit made up 13.7% of citations in Steady Demand’s 2026 study, though only about 2% in Yext’s. In Thailand, Pantip plays a similar role for many local questions. The rule is simple: answer as a helpful expert, say who you are, and never post fake recommendations. Forum users and moderators spot astroturfing fast, and a thread calling you out is the last thing you want an AI to read.

Local AEO works on three clocks. Technical fixes are fast, profile and listing cleanups take weeks, and reputation work takes months. Start with the fast, cheap items, because they make everything after them count.

Day 1. Check that robots.txt and your firewall allow OAI-SearchBot, Googlebot and Bingbot. OpenAI says robots.txt changes take about 24 hours to apply for ChatGPT search. Make sure your key pages are indexed in Google and Bing.

Week 1. Run your prompt audit to get a baseline: 15 to 30 questions, four platforms, three runs each, from your service area.

Weeks 2 to 4. Fix your Google Business Profile and your master record, then clean up Yelp, Facebook, Apple Maps, Bing Places and your industry directories. In Thailand, add Wongnai where it fits and your LINE OA details.

Weeks 2 to 8. Rewrite or create one page per main service, answer-first, with real prices, hours and areas. Add LocalBusiness schema that matches.

Months 2 to 6 and beyond. Build steady reviews with replies, and earn mentions on the lists, local media, associations and forums your audit showed AI trusts.

Every month. Re-run the prompts, compare mention rate and share of voice, and adjust.

How soon will Paws & Bubbles show up more often? Nobody can promise a date, and anyone who does is guessing. AI answers vary by run, location and platform, and the platforms keep changing. What we can say is that the businesses AI recommends tend to share the same traits: complete profiles, consistent details, strong recent reviews, clear pages and real mentions around the web. Each step above moves you closer to that picture.

If you’d like a second pair of eyes, our AEO service starts with exactly this kind of local audit across ChatGPT, Gemini and Google AI Mode, and turns the findings into a prioritized fix list.

1. SOCi: In AI-driven discovery, few brands are chosen, most disappear (2026 Local Visibility Index, 28 Jan 2026)
2. Search Engine Land: AI local visibility is up to 30x harder than ranking in Google (SOCi report coverage, 2026)
3. BrightLocal: Local Consumer Review Survey 2026 (11 Feb 2026)
4. BrightLocal: Local AI Visibility Study (16 Sep 2026)
5. BrightLocal: Uncovering ChatGPT search sources for local businesses (Dec 2024)
6. Yext Research: AI citations, user locations and query context (9 Oct 2025)
7. Whitespark: 2026 Local Search Ranking Factors (6 Nov 2025)
8. Whitespark: Case study, the prevalence of AI Overviews in local search (12 May 2025)
9. Google Business Profile Help: Tips to improve your local ranking on Google
10. Google Business Profile Help: Guidelines for representing your business on Google
11. Google Maps Help: Prohibited and restricted content (fake engagement)
12. Google Search Central: AI features and your website (updated 10 Dec 2025)
13. Google Search Central: Local business structured data (updated 8 Sep 2026)
14. Google Search Central: Review snippet structured data (updated 8 Sep 2026)
15. Google: How we're reimagining Maps with Gemini, Ask Maps (12 Mar 2026)
16. Google Maps Help: Ask questions in Google Maps
17. Google Search Help: Use AI Mode to check local availability and pricing
18. Google: AI Mode in Search gets new agentic features and expands globally (21 Aug 2025)
19. Google Thailand Blog: AI Mode now available in Thai (8 Oct 2025)
20. OpenAI Help Center: ChatGPT search
21. OpenAI: Overview of OpenAI crawlers
22. PPC Land: ChatGPT quietly rolled out location sharing (28 Mar 2026)
23. Search Engine Roundtable: ChatGPT enables location sharing for more localized near me results (31 Mar 2026)
24. Search Engine Land: How does ChatGPT conduct local searches? (Damian Rollison, SOCi)
25. Yelp: Full-year 2025 results, agreement with OpenAI (12 Feb 2026)
26. PinMeTo: ChatGPT gains Yelp reviews and ratings for local answers (24 Jul 2026)
27. PinMeTo: Gemini's local AI search cites business websites most (Steady Demand AI Citation Ledger, 20 Aug 2026)
28. Maginative: Perplexity adds Yelp data to its local search (12 Mar 2024)
29. Search Engine Roundtable: Google Ask Maps fully available in US and India (1 Apr 2026)
30. Glenn Gabe (GSQI): ChatGPT search new features, maps and local listings (Dec 2024, updated Feb 2025)
31. Relevant Audience: LINE Thailand puts 56 million users and 12 billion Official Account conversations behind its 2026 business stack (12 Aug 2026)

Keep reading

• Measurement & localHow to Measure AI Visibility and Track AI Citations
• AEO basicsWhat Is AEO? Answer Engine Optimization Explained
• AI platformsHow to Optimize for Google AI Mode
• AI platformsHow to Get Cited by ChatGPT