Winning Trades Businesses Run on AI

The trades businesses that win run their offices on AI. The clearest proof arrived this fall: KKR is reportedly paying about $2 billion for A1 Garage Door Service, a company that uses AI to catch the leads that slip. That's roughly 20 times earnings, in an industry where small shops typically sell for two to six times.

Tommy Mello founded A1 in Phoenix in 2007. On Sept. 2, Reuters reported, citing people familiar with the matter, that KKR agreed to buy it for around $2 billion. A month later, PitchBook put the price at about 20 times EBITDA and noted that the mom-and-pop garage door shops that make up most of the industry trade at 2 to 6 times. Same trade, same trucks, same garage doors. The gap is how the business runs.

TL;DR

• The trades businesses that win run their offices on AI and their jobs on people. A1 Garage Door Service is the proof, with a reported $2 billion sale to KKR at about 20 times earnings.
• AI answers the after-hours call, follows up on every lead and estimate, books the visit and asks for the review, and people keep the diagnosis and the hard conversations.
• You don't need 33 markets to copy it. The multiple goes to the shop whose leads never sit, and that's a system any owner can build.

Key fact: Small garage door companies typically sell for 2 to 6 times EBITDA. A1's reported price is about 20 times, according to PitchBook.

What happened with A1 Garage Door Service?

A1 grew from one Phoenix shop into a residential garage door company serving 33 U.S. markets, according to Qualified Remodeler, with at least 17 add-on acquisitions after Cortec Group invested in December 2022. It closed two more deals in August, the second a week before the KKR report.

It's the largest deal we can find anywhere in residential garage door services, more than double the $800 million Oak Hill Capital paid for Guild Garage Group earlier this year, at about 18 times EBITDA. One trade, two deals, and the higher multiple went to the one whose AI front office is on the record.

Buyers don't pay 20 times earnings for trucks and technicians. They pay for a machine that turns a ringing phone into a booked job the same way in every market, without the owner on the line.

How does A1 use AI?

On the office side, everywhere a lead could slip.

A1 uses ServiceTitan's AI to find calls that should have become bookings and didn't. The software listens to call recordings and flags the missed ones so someone calls back while the customer is still deciding. "It's using AI to identify (the missed opportunity)," Chase Mitchell, A1's director of performance marketing, told ServiceTitan in a January 2025 customer story. "We don't have to listen to that call."

A1 also used Chiirp, an AI system built for the trades that answers and texts new leads within seconds and keeps following up on unsold estimates, according to Chiirp founder Ryan Fenn. Mello invested in Chiirp in January 2024.

The pattern is the point. A1 didn't hand AI the garage doors. It handed AI the moments a lead goes cold: the call that ended without a booking, the new lead nobody answered in time, the estimate nobody chased.

Why does AI matter more in the trades now?

Because skilled hands are the scarce part. Mike Rowe put the shortage in four words on a Ford Pro Accelerate panel on Sept. 30: "Five out, two in." By his count, two new tradespeople replace every five who retire. A Sept. 30 national report from Jobs for the Future found formal training pathways prepare roughly 55 people for every 100 skilled trades workers the country needs.

When techs are that hard to hire, a tech's hour is the most valuable thing a shop owns. Every hour one spends returning calls or typing invoices is an hour off a job.

Will AI replace electricians, plumbers or HVAC techs?

No. Plan for the opposite. The Bureau of Labor Statistics projects employment growth from 2025 to 2035 of 11% for HVAC mechanics, 9% for electricians and 7% for plumbers, all much faster than average.

The work is physical and every house is different. AI runs the desk so the people who can do that work spend their day doing it.

What should AI do inside a trades business?

Guard every lead and every skilled hour. The split we'd make:

Homeowners are ready for it. In a survey of 1,040 U.S. homeowners by field-service software maker Housecall Pro, nearly all said speed influences who they hire, and 53% said they're comfortable with AI handling the first inquiry.

We've seen it pay off in our own builds. 4M Lawn Care used to drive to every commercial property to write an estimate and invoiced by hand each month. The system we built prices a property from its address, using county parcel data and satellite imagery read by AI, and invoices per-cut accounts the day the cut happens. At Celebrity Closet Guys, turning a customer conversation into a priced closet design took about two hours; with the AI-assisted designer we built, it takes about ten minutes. We covered the owner's side in our Sept. 24 piece.

What should you never automate?

Keep these with people:

• The diagnosis. AI can gather symptoms and photos. A licensed tech decides what's wrong.
• Anything that sounds dangerous. A gas smell or sparks at the panel go straight to a person.
• A price that needs eyes on the job. If the bid depends on what's behind the wall, it waits for a walk-through.
• The apology. When a job goes wrong, the customer should hear from the owner.

What are the honest limits?

AI gets things wrong with confidence. In the 4M build, the model reading an aerial photo will occasionally mistake a large house on a small lot for a commercial property, so the system checks parcel acreage and flags shaky bids for a walk. Anything that answers for your business needs a check like that.

Not every customer wants a bot. If 53% are comfortable with AI on first contact, nearly half didn't say so, so keep a person one tap away.

And AI won't build A1 for you. Mello spent close to two decades on training, pricing, marketing and acquisitions. AI made that machine faster. It didn't replace it, and it won't quote from a price book that lives only in your head, so write the rules down first.

Where is this headed?

Our predictions, with dates attached:

• By the end of 2027, a reply within minutes at any hour is the price of entry. Homeowners will hire whoever answers first, because nearly every homeowner in the Housecall Pro survey says speed shapes who they hire.
• By 2028, buyers ask to see the office, not just the books. We expect anyone acquiring a trades business to ask how leads are answered, followed up and booked, and to pay more for a shop where the answer is a system.
• By 2028, the office manager runs the AI front desk. The best ones will spend their days reviewing what the AI booked and taking the calls it hands off, and they'll be worth more for it.
• Through 2030, a tech's hour keeps getting more expensive. With training producing about 55 workers for every 100 needed, we expect pay for skilled techs to keep climbing all decade.

What can you do this month?

Run a missed-call audit. Pull the last 30 days from your phone system and count every call that rang out or went to voicemail, especially after 5 o'clock and on weekends. Then count the estimates you sent that never got a second follow-up. Those two numbers are what an AI front desk is worth to you before you buy one.

Then use what you already pay for. If your field-service software can answer calls, chase estimates and book into your real calendar, turn that on first, and build custom where your pricing or process sets you apart, like 4M's address-based bids.

If that's your shop, we can help you put AI on the office side of your business.