A plumber in Portland told me last month that his phone stopped ringing. Not gradually. It just stopped. His Google Business Profile still showed him in the local 3-pack for his main keywords. His reviews were solid, 4.7 stars with over 200 ratings. His website looked fine. By every traditional local SEO metric, he was doing everything right.
The problem was that people had stopped clicking. They were asking ChatGPT for a plumber recommendation instead.
This is the story playing out in thousands of small markets right now, and it is the single biggest shift in local search since Google introduced the local pack back in 2014. AI search is not replacing Google Maps overnight, but it is quietly rerouting how people discover and choose local businesses. And the rules that govern which businesses get recommended by these AI systems are nothing like the rules we spent a decade mastering.
The numbers that should worry you (and the ones that should excite you)
SOCi's 2026 Local Visibility Index analyzed performance data from nearly 350,000 locations across 2,751 multi-location brands. The findings are jarring. Only 1.2% of locations were recommended by ChatGPT. Gemini recommended 11% of locations. Perplexity landed in the middle at 7.4%. Compare those numbers to Google's local 3-pack, where brands appeared 35.9% of the time.
Put another way, getting your business recommended by an AI assistant is somewhere between three and thirty times harder than ranking in traditional local search. That gap is enormous, and it is not closing quickly.
But here is the thing that should excite you: because so few businesses are being recommended, the ones that figure this out early own the conversation. When ChatGPT tells someone to call your plumbing company, there is no list of ten blue links competing for attention. There is no ad above you. It is just your name, your phone number, and a reason to call. The conversion rate on those recommendations is absurdly high compared to traditional search clicks.
The businesses that get this right in 2026 are going to look back and realize this was the cheapest customer acquisition channel they ever found. The window will not stay open forever.
Where AI search gets its local data (and why Google Maps matters more than ever)
I used to think of Google Maps as just one channel among many. Yelp, Apple Maps, Bing Places, Google Maps, whatever. You listed on all of them and moved on. That mental model is dead.
Google Maps is now one of the primary data sources for Gemini, and because Gemini powers AI Overviews in Google Search, it is also the data backbone of the biggest AI search product on the planet. ChatGPT has built integrations that pull from Google Maps data. Perplexity cross-references it. Apple's AI features lean on it for local recommendations.
Your Google Business Profile is not just a listing anymore. It is the canonical record of your business that multiple AI systems read from. The accuracy of your business information on Google Maps was 100% on Gemini, compared to roughly 68% accuracy on ChatGPT and Perplexity. That 68% number is terrifying if you think about it. One in three times an AI recommends a business through those platforms, the phone number, address, or hours might be wrong. And that creates a weird advantage: if your information is correct across all platforms while your competitors' information is messy, you get recommended more reliably.
This means the boring work of citation management, making sure your name, address, and phone number are identical everywhere, matters more in 2026 than it has in years. Not because of some outdated SEO tactic, but because AI systems are pulling from these sources and they are terrible at reconciling conflicting information.
Unstructured citations: the fourth most important factor nobody talks about
Whitespark's 2026 Local Search Ranking Factors survey, which remains the gold standard for understanding what moves the needle in local search, identified the quality and authority of unstructured citations as the fourth most important factor for AI search visibility.
Let me explain what that means in plain terms. A structured citation is your listing on Yelp or Yellow Pages or the local chamber of commerce directory. Name, address, phone, website, neatly formatted. We have been building those for years.
An unstructured citation is when a local newspaper mentions your business in an article about the best HVAC companies in the area. It is when a food blogger writes about your restaurant without any formal listing format. It is when a Reddit user in your city's subreddit recommends your auto shop by name. These mentions are messy and organic and exactly the kind of signal that AI systems treat as authentic endorsement.
Think about why this makes sense from the AI's perspective. If the only places your business appears online are structured directories, that looks like you paid to be listed. Which you did. If your business also shows up in local news articles, blog posts, government directories, and community forums, that looks like real-world reputation. AI systems are pattern matchers at their core, and the pattern of a business that exists only in paid directories looks very different from one that the community actually talks about.
I have been advising small business clients to spend less time on directory submission services and more time on local PR. Getting mentioned in a local newspaper article about "spring home improvement tips" where they quote your roofing company is worth more than fifty directory listings right now. Getting your restaurant mentioned in a local food blogger's roundup matters. Getting your law firm cited in a local business association newsletter matters.
The hard part is that you cannot automate unstructured citations the way you can submit to directories. They require actual relationships with local media, actual participation in community events, actual reasons for people to mention you. Which is, I think, exactly the point. AI systems are rewarding the businesses that are genuinely embedded in their communities.
Review velocity beats review count, and it is not even close
I want to be direct about something that I think a lot of local SEO advice gets wrong. Having 500 reviews does not matter nearly as much as getting five new reviews every week.
Google's algorithm, and increasingly the AI systems that feed from Google's data, responds to review velocity. That is the rate at which new reviews appear over time. Not spikes, not historical totals, but consistency. A steady stream of reviews signals three things: the business is actively serving customers, the service quality is consistent, and the business is operational and relevant right now.
I have seen businesses with 150 reviews outrank businesses with 800 reviews because the smaller business was getting three to four new reviews per week while the larger business had not received a review in two months. Google interprets that silence as a signal. Maybe the business slowed down. Maybe quality dropped. Maybe they are not relevant anymore. Whatever the interpretation, stale review profiles perform worse than active ones.
For AI systems, review recency matters even more because these systems are explicitly designed to surface current, reliable information. When Gemini or ChatGPT pulls data about a local business, the recency of reviews factors into whether they recommend it. A business with a 4.8 rating from 2024 reviews looks less trustworthy than a business with a 4.5 rating from reviews posted last week.
The practical advice here is boring but effective. Build a review request process into your business operations. Every completed job, every sale, every service call should include a follow-up asking for a review. Not a batch email blast once a quarter. A steady, automated, but personalized drip. The goal is rhythm, not volume.
The search behavior shift that changes everything
Here is what I think most local business owners are missing about AI search: it is not just a new channel. It is a fundamentally different way people make decisions.
When someone searches Google for "plumber near me," they get a list. They call two or three. They compare prices. The plumber with the best combination of proximity, reviews, and pricing wins.
When someone asks ChatGPT or Perplexity "I have a leaking pipe under my kitchen sink, who should I call in Portland?", they get one or two recommendations with context about why those businesses are good. They call one. The decision is already made before they pick up the phone.
This collapse of the consideration phase is massive. In traditional search, you needed to be visible and then persuasive. Your listing needed to catch their eye, your reviews needed to reassure them, your website needed to close the deal. With AI recommendations, the AI handles the persuasion. It has already evaluated you against alternatives and decided you are the right answer.
The businesses I work with that are getting AI recommendations report dramatically higher conversion rates from those leads. These are not people shopping around. These are people who were told by an AI they trust that this specific business is the one to call. That level of pre-qualification used to require expensive paid advertising funnels.
But the flip side is terrifying. If you are not the business being recommended, you are invisible. There is no second page to scroll to. There is no map to expand. The AI picked someone else, and you do not exist in that interaction.
What actually works: a practical framework for AI local visibility
I am going to be honest that nobody has a perfect playbook for this yet. The field is too new, the AI platforms change too quickly, and the data is still accumulating. But based on what I have seen working with local businesses over the past year, here is what I would prioritize.
First, your Google Business Profile needs to be immaculate. Not good. Not fine. Perfect. Every field filled out. Photos updated monthly. Products and services listed with genuine descriptions, not keyword-stuffed nonsense. Business hours accurate to the minute. Q&A section actively managed. This is your canonical data source for every AI system, and any gap or inaccuracy propagates across all of them.
Second, invest in content that explains what you do, where you do it, and who you serve, written in natural language that an AI can extract information from. I am not talking about location pages stuffed with "plumber in [city name]" repeated thirty times. I am talking about genuine service descriptions that answer the kind of questions people ask AI assistants. "Do you handle slab leaks?" "What areas do you serve?" "Are you available on weekends?" When someone asks ChatGPT a specific question about a local service, the AI needs to find the answer in your content.
Third, build your unstructured citation profile. Get mentioned in local media. Participate in community events that generate press. Write guest posts for local business blogs. Get quoted in industry roundups. Every organic mention of your business in a context that is not a paid directory strengthens your AI visibility.
Fourth, maintain review velocity. I said it already but it bears repeating because so many businesses treat reviews as a one-time project rather than an ongoing operation. The businesses that win at AI local visibility in 2026 are the ones getting a consistent stream of recent, authentic reviews.
Fifth, and this one surprises people, be present on Reddit and local community forums. Reddit is the second most cited source for both Google AI Overviews and Perplexity, and local subreddits are exactly where people discuss and recommend local businesses. You do not need to spam your business name in every thread. You need to be a helpful participant in your local community online, so that when someone asks "who is the best electrician in Austin?" your name comes up naturally.
The uncomfortable truth about the timeline
I do not want to create false urgency, but I also do not want to be dishonest about what I am seeing. The shift from traditional local search to AI-mediated local discovery is happening faster than most people expect. Calls from Google Business Profiles are declining even for businesses that rank well, because users are getting answers before they ever interact with a listing.
The businesses that adapt now, while AI visibility is still relatively uncrowded, will have a significant advantage. The 1.2% recommendation rate on ChatGPT is not going to stay that low forever. As AI systems get better at local data and more businesses optimize for AI visibility, the competitive landscape will tighten. But right now, the bar is low enough that a well-optimized small business can punch way above its weight.
I have worked in local SEO for over a decade, and this is the most significant shift I have seen since the arrival of mobile search. The businesses that treated mobile as "just another screen" spent years catching up. I do not want to see the same thing happen with AI search.
The good news is that most of what works for AI local visibility also makes your business better. Having accurate information everywhere, earning genuine community mentions, maintaining a steady flow of happy customer reviews, creating content that actually answers people's questions. None of this is manipulative. None of it is a hack. It is just the stuff that good businesses should be doing anyway, finally being rewarded by the systems that connect customers to businesses.
That plumber in Portland? He restructured his Google Business Profile, started a consistent review request process, got mentioned in two local home improvement articles, and began posting helpful answers in the Portland subreddit. His phone started ringing again within six weeks. Not from the same channels as before. From better ones.