Short answer: I asked the question a panicked homeowner asks, word for word, in five markets. ChatGPT named five or six restoration companies every single time, with phone numbers attached. Google gave an actual AI answer in exactly one of the five towns. And in every town, the top pick was a local or independent company, not the national franchise. The franchise was in the room. It just never got the first slot.
Who do you call at 2 AM?
Picture the worst night of someone's year.
A supply line lets go under the upstairs sink. They wake up to a sound they do not recognize, step out of bed, and their foot lands in an inch of water.
They are not opening a laptop. They are not comparing five websites. They grab the phone and say, "Who do I call for water damage near me?"
Somebody is getting that job. The only question is whether the machine says your name.
How I ran it
I asked the customer's question, not a marketing question. Same wording every time:
Who should I call for water damage in [city].
Pensacola, Florida. Mobile, Alabama. Tampa, Florida. Nashville, Tennessee. Dallas, Texas.
No brand names in the prompt. No leading questions. In ChatGPT I ran all five in one clean conversation, logged out, exactly like a stranger would. For Google I was signed into a normal account, which pushed an ads panel over the top of the page. Then I wrote down every name that came back, in the order it came back.
What Google did
Almost nothing.
One town out of five gave me a real AI answer. Mobile, Alabama. Here is the whole thing:
"You can call Roto-Rooter Plumbers & Water Clean-up at 251-479-9454 or SERVPRO of Mobile County at 251-633-7500 for 24/7 emergency water damage cleanup and restoration in Mobile, Alabama."
Two companies. Two phone numbers. That is the complete answer Google gave.
The other four towns got no AI answer at all. Just links. In Pensacola the top of the page was a paid listing for SERVPRO of West Pensacola and a paid text ad for VetCor Water Restoration. In Tampa, Nashville and Dallas there was no local map pack either, only ordinary web results and a few Yelp snippets.
So if you have been telling yourself the AI thing is overblown because you still see regular search results, this is the part that feels reassuring.
Do not get comfortable. Keep reading.
What ChatGPT did
Every town. Every single one.
Pensacola gave me five companies with five phone numbers. Mobile gave five, plus two more at the bottom. Tampa gave six. Nashville gave six plus three. Dallas gave six plus three.
Nobody got refused. Nobody got told "I don't have local information." The machine had an answer ready in every market I asked about, with a phone number for every company it named, and star ratings and review counts on a good number of them.
Now here is the part I did not expect. In all five towns, the first company named was a local or independent outfit.
Pensacola started with Paul Davis Restoration of the Emerald Coast, 4.9 stars on 181 reviews. Tampa started with Big Wave Restoration, 5.0 on 384. Dallas started with 24 Hour Flood Pros at 5.0 on 1,115. Mobile started with United Water Restoration Group of Mobile, 5.0 on 23 reviews. Nashville started with Blue Chip Restoration.
And SERVPRO?
In the answer every time. Never first. Second in Pensacola, Mobile and Tampa. Third in Nashville. Fifth in Dallas, where the machine put four local companies above it.
Read that again. In Nashville, SERVPRO had 4.9 stars and 419 reviews, more reviews than anything else in the list, and it still landed third.
So ratings do not matter?
They matter. They are just not the whole thing, and anybody selling you "get more reviews and you win AI" is selling you half a sentence.
Look at Mobile. PuroClean came third with 4.7 stars and 50 reviews, ahead of a 4.9 and a 4.8. Look at Pensacola, where the 4.9 with 181 reviews beat a 5.0 with 24. Dallas was the one clean story, with the top four all sitting at 5.0 and the two 4.7s at the bottom.
What I see in that is not a rating formula. It is a machine trying to answer a question and reaching for whatever it can describe with confidence. A big review count tells it you are real and busy. A rating tells it people were happy. Neither one tells it that you handle a flooded basement on a Sunday.
That has to come from somewhere else. Your profile. Your services. Your hours. The words people use when they write about you.
Why the franchise kept showing up but never won
This is the most useful thing I found, and I think it is the thing nobody is talking about yet.
The national brand was in every answer. If you are a franchise owner reading this, that is good news. The machine knows who you are, it knows your local franchise by name, and it will put you on the list.
But it put a local company first in every single town. And it said so out loud. In Tampa it told me it was giving me options "rather than just picking a national franchise." In Nashville the phrase was "rather than just the biggest name."
That is the machine making a judgment about what a person in trouble actually wants. Not the biggest logo. The one that feels closest and most specific.
There is a lesson in there for every franchise owner who has spent years paying for a brand name and hoping it carries them. The brand gets you in the door of the answer. It does not get you the first line.
What this means if you are the guy on the list
Being on the list is not the same as getting the call.
Think about what you actually do with a list of six companies and six phone numbers. You call the first one. If they do not pick up, maybe you call the second. You are standing in an inch of water. You are not working your way to number five.
So position one is worth most of the value, and position five is worth almost none. The list is not the win. The first line is the win.
What this means if you are not on the list
Every town I tested had a full list. Five or six companies, phone numbers, ratings.
Which means in these five markets, the answer is already crowded. There is no empty seat waiting for you to show up with a better website. You have to displace somebody.
That is a different job than ranking in search, where there are ten spots and you can creep up one at a time. Here there are five or six, and the machine is confident about them. You get in by being more describable than the company it currently names.
Three things to do this week
First, run the test yourself. Not in a browser where you are signed into your own Google or your own business account. Open a private window and ask the question the way a stranger would, with no brand name in it. Ask it in ChatGPT too, in a fresh conversation. Write down who got named and where you landed.
Second, make your profile say what you actually do. Every service. Real hours, including the overnight ones. Your real service area. If your crew rolls at 2 AM and nothing on the internet says so, no machine can put you in the 2 AM answer.
Third, fix your reviews at the source. Not the count. The content. If every review says "great company, five stars," you have given the machine nothing to describe. The reviews that get you picked talk about the basement, the weekend, the hour it took to arrive, the drying equipment, the insurance call. Ask your customers to mention what you actually did for them.
That is not gaming anything. That is being legible.
If you want to see how your reviews and your listings actually look across every location side by side, that is the exact problem we built Were You Happy to solve. It scores each location and shows you the weak one before a customer does.
One honest caveat
This is one run, on one day, in one conversation, from one part of the country. ChatGPT was logged out, Google was not.
Your results will differ. Mine might differ tomorrow. AI answers move, and they move for reasons none of us can see.
Which is exactly why you check instead of assuming. Ten minutes a month, in a private window, in the towns you actually serve, tells you more than any article about AI search ever will.
Including this one.
FAQ
Does ChatGPT recommend local businesses?
Yes, and specifically. In five markets it named five or six local restoration companies per town, each with a phone number, and most with star ratings and review counts. It did not send anyone to a generic national directory.
Why did Google show an AI answer in only one city?
AI answers are not live everywhere yet. It depends on the market, the question and how confident the system is that it can give a good local answer. Google showed a "Thinking" indicator in four of my five searches and then settled into plain links.
Does SERVPRO show up in AI answers?
In my test, yes, in every single market, always under the local franchise name. It was never the first company named. The top pick was a local or independent company in all five towns.
Do more reviews get you picked first by AI?
Not by themselves. In my test, one company with 419 reviews and 4.9 stars was named third behind two companies with no rating shown. Another with 4.7 and 50 reviews came in ahead of a 4.9 and a 4.8. Reviews clearly matter. They are not the only input.
How do I find out what AI says about my business?
Open a private browser window. Ask "who should I call for [your service] in [your city]" with no brand name in the question. Then ask the same thing in ChatGPT in a fresh conversation. Write down the names you get and where you land.
Can I pay to be the name AI recommends?
Not the organic answer. In Pensacola, the closest thing to a guarantee I saw was a paid listing for a local SERVPRO and a paid text ad for another company. That is a rental. The named answer is the one that compounds.
Want me to check your market?
Run it yourself tonight. Private window, no brand name, the way a customer would ask it.
If you would rather have a baseline than a guess, you can score your locations free at wereyouhappy.com and see where you actually stand before you start fixing anything.
If it does not say your name, send me your city on LinkedIn and I will run it and tell you exactly what I see. No pitch. Just what the machine says.
.png)
.png)