An HVAC contractor running a four-man crew missed three calls one Tuesday morning. His team was on a commercial roof pulling an RTU. The office phone rang at 8:12, 8:41, and 9:03. All three went to voicemail. Two of those callers needed full system replacements — jobs worth $7,000 to $12,000 each. By 9:15, both had called competitors.
That is not a story about artificial intelligence replacing his trade. It is a story about revenue walking out the door because nobody picked up the phone.
Read the headlines, though, and you would think every contractor in America should be worried about a robot showing up with a torch and a condenser coil. The panic is loud. The data tells a different story entirely.
What the research actually measured
The Anthropic Economic Index studied AI exposure across hundreds of occupations. Exposure, in this context, means the share of tasks within a job that large language models can meaningfully perform or assist with. Not replace. Perform or assist.
The results for trades were striking in how low they landed.
Tradespeople as a broad category sit at roughly 2% AI task exposure. Electricians come in at 7%. HVAC technicians at 20%. Plumbers at 22%. Compare that to computer programmers at approximately 74.5%.
Read those numbers one more time. A plumber's daily work overlaps with what AI can do by about one-fifth. A programmer's overlaps by roughly three-quarters. If anyone should be reading the room, it is the people writing the software, not the people sweating copper.
The 20% figure for HVAC does not mean a fifth of technicians lose their jobs. It means a fifth of the tasks in the role — things like documentation, lookup, reporting — have some overlap with what a language model can handle. Brazing a lineset is not on that list. Pulling a vacuum is not on that list. Diagnosing a failing TXV by feel and sound and pressure readings is not on that list.
The labor picture reinforces it
The U.S. Bureau of Labor Statistics projects roughly 81,000 electrician openings per year through 2034. That figure accounts for growth, retirements, and attrition. The pipeline of new apprentices is not keeping pace.
Electrification is accelerating the gap. Heat pump installations, EV charger rough-ins, panel upgrades from 100-amp to 200-amp service — the volume of skilled electrical work is climbing while the labor pool stays flat or shrinks. Similar dynamics play out across HVAC, plumbing, and general contracting.
The problem facing trades is not displacement. It is a shortage of hands. AI is not going to show up at a job site and thread black iron gas pipe. The country needs more tradespeople, not fewer.
Where the real threat lives
Here is the part worth losing sleep over, and it has nothing to do with machine learning.
BrightLocal reported in 2024 that 62% of calls to home-service businesses go unanswered. Invoca's 2023 study found 85% of callers who reach voicemail will not call back. And an industry aggregate from 2024 puts the number at 67% of those callers contacting a competitor within five minutes.
Sixty-two percent unanswered. Eighty-five percent gone for good. Two-thirds calling your competitor before you finish the job you are on.
That is the threat. Not a language model. The phone.
A four-man HVAC crew running two vans cannot answer every call during a heat wave. A solo plumber crawling under a pier-and-beam foundation at 2pm is not pulling out a phone. An electrician on a ladder with a hot panel open should not be reaching for anything except a meter.
The nature of trade work means the phone rings when you cannot pick it up. That is structural. It does not get fixed by hiring faster or working harder.
The version of AI worth thinking about
There is a useful distinction between two layers of AI in a service business.
The cognitive layer is the part that does the skilled work — diagnosing, designing, installing, repairing. That layer barely overlaps with what AI can do in trades. The Anthropic data confirms it.
The operational layer is the part that handles scheduling, call intake, customer communication, and follow-up. That layer has significant overlap with what AI handles well right now. Answering a phone call, capturing a caller's name and address, booking a time slot, sending a confirmation text — none of that requires a journeyman card.
The contractors who benefit most from AI are not the ones threatened by it. They are the ones who use it to stop losing calls while their crews are on the job.
How DoubleXL Answer works
DoubleXL Answer is an AI phone answering service built for service businesses. It does three things:
- Answers the call. Every ring, every time — during jobs, after hours, weekends, holidays.
- Books the customer. Captures details, schedules according to your availability, handles the intake you would do yourself if you were sitting at the desk.
- Texts the summary. You and the caller both get a text with the details so nothing falls through.
Starting at $99/mo. No contracts, no per-minute billing surprises.
The business case plainly stated
Replacing a tradesperson with AI is not realistic and not happening. Replacing the voicemail that is costing you thousands every month in missed work — that is happening right now, for less than the cost of a single service call.
If you are an HVAC contractor, plumber, or electrician reading headlines about AI taking your job, set those aside. Your trade is safe. Your phone is the problem.
Run the numbers yourself with the missed-call calculator, or read more about how AI fits into the trades workforce picture at /workforce-ai.
Questions: hello@double-xl.com | +1 (385) 396-4901
Frequently Asked Questions
Will AI replace electricians and HVAC techs in the next few years?
No. The Anthropic Economic Index puts trades at 2% AI task exposure overall, with electricians at 7% and HVAC techs at 20%. Those numbers measure task overlap, not job replacement. The physical, diagnostic, and judgment-heavy nature of trade work has almost no overlap with what current AI systems can do. The BLS projects 81,000 electrician openings per year through 2034 — demand is growing, not shrinking.
Why do home-service businesses lose customers even with steady demand?
The phone. BrightLocal's 2024 data shows 62% of home-service calls go unanswered. Invoca found 85% of those callers never try again. And 67% contact a competitor within five minutes (industry aggregate, 2024). Demand is not the issue. Answering the demand is. When your crew is on a roof or under a house, the calls that come in go to voicemail — and voicemail is where revenue dies.
What does DoubleXL Answer actually do for a service business?
It answers your phone when you cannot. Every call gets picked up — during jobs, after hours, weekends. The AI captures the caller's information, books according to your availability, and texts both you and the customer a summary. Starting at $99/mo, it covers the structural gap between when your phone rings and when your hands are free. It is not a dispatcher replacement. It is the layer that keeps your pipeline full while your crew works.