Say you run a two-truck HVAC shop in Chandler, Arizona. Call it Summit Heating & Air. Tonight a homeowner somewhere in your service area loses their AC, and instead of googling, they ask ChatGPT: "who's a reliable AC repair company in Chandler?"
The answer comes back in about four seconds. It names the 40-truck outfit across town and two others. Summit isn't on the list.
That outcome had little to do with luck and almost nothing to do with company size. AI assistants build recommendations from a short list of sources they can crawl and check, and most shops Summit's size barely show up in half of them. You can fix that, and for now the fix is cheap, because almost nobody in the trades has started on it.
People call this work GEO, short for generative engine optimization. SEO gets you onto a results page with ten links. GEO gets you into an answer that names three companies and leaves everyone else out.
An AI answer has no page two. If the assistant doesn't name you, the homeowner never hears of you.
What the assistant read before it answered
An assistant recommending an HVAC company leans on a few sources it trusts because it can check them.
The first is the company's own website, read as data rather than as a brochure. The 40-truck outfit's site uses schema markup to tell crawlers what it does and where it works, down to the hours it answers the phone. Summit's site says "Comfort is our business!" over a stock photo. An assistant can quote the first site in an answer. With the second it would have to guess, and it won't. It names someone else.
Reviews come next, and volume matters less than recency and specific language. "Fixed our AC" gives an assistant nothing to work with. "Diagnosed a failed capacitor and had the part on the truck" is a checkable detail, the kind an assistant repeats when someone asks who's good with older units. Every one of Summit's five-star reviews says "great service!!" That tells the assistant customers were happy. It says nothing about what Summit can fix.
Then there's consistency across the open web, meaning the same name, phone number, and service list on the website, the Google Business Profile, Yelp, and the trade directories. Suppose Summit's Google profile says "Summit Heating and Air LLC," closes at 5, and lists "AC repair." Meanwhile the website says "Summit HVAC," closes at 6, and lists "cooling services." Each mismatch makes the assistant less sure it knows who Summit is. An assistant that isn't sure leaves the business out.
Ad spend isn't on that list. Right now you can't buy your way into the organic answer. That will change at some point, which is one more reason to do this work now.
Fix one: describe your business in a format crawlers read
The best use of an afternoon in this post is adding LocalBusiness schema to your website. Summit's would look roughly like this. Swap in your own details and send it to whoever built your site.
{
"@context": "https://schema.org",
"@type": "HVACBusiness",
"name": "Summit Heating & Air",
"telephone": "+1-480-555-0142",
"url": "https://example.com",
"address": {
"@type": "PostalAddress",
"streetAddress": "1234 W Example Rd",
"addressLocality": "Chandler",
"addressRegion": "AZ",
"postalCode": "85225"
},
"areaServed": ["Chandler", "Gilbert", "Tempe", "Mesa"],
"openingHours": "Mo-Sa 07:00-18:00",
"makesOffer": [
{ "@type": "Offer", "itemOffered": { "@type": "Service", "name": "AC repair" } },
{ "@type": "Offer", "itemOffered": { "@type": "Service", "name": "Furnace repair" } },
{ "@type": "Offer", "itemOffered": { "@type": "Service", "name": "HVAC installation" } }
]
}
That block turns "Comfort is our business!" into facts an assistant can check and repeat. Without it, the assistant has to infer every one of those fields on its own. Small shops lose to big ones on inference, because the big shops leave less to infer.
Fix two: ask customers what you fixed
Change what you say after a job. Instead of asking for "a quick review," ask one question: "Would you mention what we fixed?" Ten reviews like "replaced the blower motor same day" are worth more than a hundred that say "great service!!" Assistants look for proof of the work. Specific reviews let a two-truck shop out-credential a 40-truck one on the jobs it does best.
We hold our own marketing to the same standard. Our case studies show the math behind every number so a skeptical buyer can check it.
Fix three: the same details everywhere
Settle on one official name, phone number, set of hours, and service list. Then make every listing match, including the site, the Google Business Profile, Yelp, Nextdoor, and any directory that mentions you. It's tedious. It also decides whether an assistant vouches for Summit or hedges and names somebody else.
Once a month, ask the homeowner's question again. Put it to ChatGPT, Perplexity, and Google's AI mode for each city you serve, and screenshot what comes back. That folder of screenshots becomes your GEO scoreboard, and it's more honest than any ranking report.
FAQ
Do people hire the businesses an AI recommends?
A growing share do, and they behave differently from search traffic. Someone who gets two names from an assistant tends to call both and decide fast. They rarely shop around after that, so you get fewer leads and warmer ones. The channel is small next to Google today, but it keeps growing, and an early position costs little.
Is GEO just SEO with a new name?
About half of it overlaps. Clean site structure and consistent business data help with both, and so does real content. GEO adds a preference for being quotable. Assistants favor sources that state checkable facts plainly over pages tuned to rank, and a page written to be cited reads differently from one written to be clicked.
Can I pay someone to "do GEO" for me?
You can. Most of what's sold under that label is the three fixes in this post at agency prices, so do those first. Outside help earns its fee on the deeper layer, such as keeping entity data consistent at scale or adding structured data past the basics. Content built around the questions people put to assistants belongs in that layer too.