Getting New Customers in 2026. What Changed and What Actually Works Now

There’s a shift happening in how people find businesses, and most business owners haven’t noticed it yet.

A couple of years ago, someone looking for a dentist, a plumber, or a real estate agent would open Google, type in a search, and scroll through a page of results. Ten blue links. Maybe a map with some pins. The business owner’s job was to be somewhere on that page.

That still works. But something else started happening alongside it. People began asking ChatGPT and Gemini for recommendations. Not as a novelty, but as a real way to decide where to spend their money.

“Best dentist in San Francisco.” “Emergency plumber Austin.” “Find me a good realtor in Miami.” Casual questions typed into a phone. And instead of a page of links to sort through, they get a direct answer. Two or three names, presented with confidence, as if a knowledgeable friend just gave them advice.

The businesses whose names come back in those answers are getting customers from a channel that didn’t exist two years ago. The businesses whose names don’t come back don’t even know those customers existed.

How ChatGPT and Gemini Decide Who to Recommend

We spent the past several months tracking how ChatGPT and Gemini recommend local businesses. We tested real queries across seven industries and six cities. The results revealed patterns that most business owners would find surprising.

The first surprise was how much the answers vary. A dental practice in San Francisco that was recommended every time on one platform didn’t appear once on another. A plumbing company in Austin that dominated three platforms was completely invisible on ChatGPT, the most popular one. The biggest brand name in American plumbing (Roto-Rooter) couldn’t get a single ChatGPT mention while smaller local companies owned the recommendations.

The second surprise was how wrong a single check can be. A barbershop owner in Boston who opens ChatGPT and sees their shop recommended might think everything is fine. But on Gemini, a completely different shop is the top pick. On Grok, a third one. Three platforms, three different winners, each absolutely confident in its answer.

ChatGPT Recommends Businesses It Can Describe with Facts

After looking at hundreds of recommendations, one pattern stood out above everything else. The businesses that showed up consistently weren’t always the best reviewed, the oldest, or the most advertised. They were the ones that gave AI the most to work with.

“Family-owned since 2011, serves the Marina District, specializes in emergency repairs, 4.8 stars from 180 reviews.” That’s a recommendation ChatGPT can make with confidence. “We provide excellent service in a comfortable environment” is a sentence that appears on a thousand websites and gives AI no reason to pick one over another.

This showed up across every industry. Law firms in Chicago with detailed case results, named specialties, and factual bios got recommended. Firms with generic “committed to excellence” language didn’t. Realtors in Miami with specific neighborhood expertise, transaction histories, and market knowledge got recommended. Agents with a headshot and a phone number didn’t.

The businesses getting recommended treat their online presence like a resume. Specific, factual, verifiable. Not marketing copy.

Why a Google Maps Profile Is No Longer Enough

The most common objection we hear is “I already have a Google Maps profile, I get customers from it, everything’s fine.”

Google Maps still matters. It’s still probably the most important single listing for a local business. But a Google Maps profile alone is no longer enough.

Google’s data isn’t fully accessible to other platforms. ChatGPT can’t read Google Maps reviews the same way Gemini can (Gemini is Google’s own product). Perplexity has its own search index. Grok pulls from yet another set of sources.

A business with 200 glowing Google reviews and a polished Google Business Profile might still be invisible on ChatGPT because ChatGPT draws from different sources. We saw this repeatedly. An emergency plumber in Austin with hundreds of Google reviews showing zero results on ChatGPT. A school that parents rave about on Google Reviews not mentioned once by three out of four platforms.

The businesses that showed up everywhere had presence beyond Google. Yelp, industry directories, local “best of” articles, mentions in neighborhood guides. Each additional independent source that confirms who the business is made the recommendation more likely.

AI Doesn’t Always Know Who It’s Recommending

Beyond the platform disagreements, we found something more fundamental. AI frequently can’t tell the difference between a business and the person who runs it.

With realtors in Miami, some platforms recommended individual agents by name while others recommended the brokerages they work for. A buyer asking Gemini would get “Dora Puig” and know exactly who to call. A buyer asking ChatGPT would get “Compass,” a company with thousands of agents nationwide. That’s like recommending “Marriott” when someone asks for a good hotel.

The same confusion hit every category. A dental practice listed as “Dr. Robert Soto” on one platform and “Robert Soto DDS” on another. AI treats each version as a separate business. One barbershop in Boston appeared under three different names across platforms. Same shop, three entries, visibility fragmented instead of adding up.

The fix is tedious but free. Make sure the business name matches everywhere it appears online.

Why None of This Shows Up in Google Analytics

Here’s what makes this tricky. When a customer asks ChatGPT for a recommendation, visits that business’s website, and books an appointment, the business sees a new visitor in their analytics. Maybe the referrer says “chatgpt.com.” More often it doesn’t say anything useful at all.

The business has no way of knowing how many potential customers asked ChatGPT and went to a competitor instead. There’s no “impressions” metric. No “position #7” to optimize toward. The customer either heard the business name or didn’t. And if they didn’t, the business never finds out. The competitor down the street might be getting five new inquiries a month from ChatGPT, and there would be no way to know.

This is fundamentally different from Google, where a business can track everything. With ChatGPT and Gemini, the entire process is invisible unless someone actively checks. Which is why most business owners don’t realize this is happening at all.

What a Missed AI Recommendation Actually Costs

The financial weight of a missed recommendation depends on the industry, but it’s real in every one of them.

A missed recommendation for a plumber at 2am means losing an emergency call that could have turned into a $10,000 customer relationship over the next few years. A missed recommendation for a private school in Los Angeles means a family that never scheduled a campus tour. At $40,000 a year in tuition, that one family represents hundreds of thousands in lifetime revenue. A lawyer not showing up on ChatGPT in Chicago could be losing $50,000 cases. Even a barbershop invisible on Gemini in Boston is losing $5,000 over the next few years of haircuts from one customer.

The math gets uncomfortable fast for high-ticket businesses. We tracked Italian kitchen showrooms in Miami where a single project runs $50,000 or more. If ChatGPT steers even a handful of customers away from a showroom and toward a competitor, that’s hundreds of thousands of dollars in missed revenue per year.

And the businesses losing these customers have no idea it’s happening, because the loss is completely invisible.

What a Business Owner Can Actually Do About It

Everything above might sound technical, but the actions that actually move the needle are things any business owner can do without hiring anyone.

Check how the business name appears online. Search for the business on Google, Yelp, industry directories, and the business’s own website. If the name, phone number, or address is different in any of those places, fix it. AI treats every variation as a separate business and splits the visibility between them. This takes an afternoon and costs nothing.

Add specific facts to the website. Most business websites read like marketing brochures. “Quality service, experienced team, customer satisfaction.” None of this helps AI recommend the business over anyone else. What helps is specifics. Year founded. Number of employees. Languages spoken. Insurance plans accepted. Neighborhoods served. Service area. Certifications. Pricing, even rough ranges, if the business is comfortable sharing it. These are the facts AI turns into recommendations. A dentist who lists “accepts Delta Dental and Cigna, Spanish-speaking staff, same-day emergency appointments available” gives ChatGPT more to work with than one who says “caring dental professionals dedicated to your smile.”

Get listed on one more platform this month. If the business only exists on Google and its own website, AI doesn’t have enough to go on. Pick one additional platform that makes sense for the industry. Yelp for restaurants and services. Niche.com for schools. Avvo or Justia for lawyers. Houzz for home services. A detailed, accurate profile on one more source can make the difference between being recommended and being skipped. The key word is detailed. A bare-bones profile with just a name and phone number doesn’t help much. Fill in the description, add photos, list services.

Ask five happy customers to leave a review somewhere other than Google. Google reviews are great for Google. But ChatGPT and Perplexity pull from other sources. A handful of Yelp reviews, or reviews on an industry-specific platform, gives AI another signal that this business is real and trusted. The easiest way to do this is to send a quick text to a loyal customer after their next visit. Most people are willing. They just need to be asked.

Check once a month. Open ChatGPT, type “best [business type] in [city]” and see what comes back. Then try the same on Gemini. It takes five minutes and the results will be different each time. We wrote a step-by-step guide to checking AI visibility that walks through the whole process. Save the results somewhere so next month there’s something to compare against.

None of these require a marketing degree or a budget. They require attention, which is something most business owners can spare for an hour a month when the payoff is new customers from a channel their competitors aren’t watching yet.

Why the Businesses That Move First Will Win

Right now, most businesses aren’t paying attention to this. Their competitors probably aren’t either. Which means the first businesses to show up in AI recommendations in any given category and city will have the space mostly to themselves.

That won’t last. As more business owners catch on, the competition for those recommendations will get crowded. The businesses that start now are building visibility that compounds over time, the same way early Google SEO paid off for the businesses that moved first.

None of the steps above require a marketing agency or a technology budget. They require attention. An afternoon to clean up listings. An hour a month to check and adjust. The kind of work that’s easy to skip when the phone is still ringing from existing channels.

But the phone rings because of channels someone built years ago. The businesses that will have full phones in 2028 are the ones building new channels right now.