Ask around about AI tools for tradespeople and you'll hear two stories. One says AI is about to change everything from estimating to diagnostics. The other says nobody on a job site has time for it. The record so far supports neither. Contractors are using AI, but mostly in the office and the sales process, while the tools aimed at the technician on a ladder are newer and less proven.
This piece sorts what trade publications have documented contractors actually doing from what is still mostly a vendor pitch, and lays out some ground rules for using it without getting burned. It's about using AI at work. If you want to know whether AI threatens trade jobs, read the trades AI can't touch.
Where AI tools for tradespeople are actually used today
A November 2025 Plumbing & Mechanical feature, "Smart tools in skilled hands", profiled contractors using AI. Nearly every use it described happens before or after the truck rolls, not on the job itself:
- Dispatch and routing. A Virginia HVAC company uses a dispatch tool called Probook, which its CEO says weighs more than 100 variables to match technicians to jobs. Another contractor described AI routing that considers location, technician skills, traffic and expected job length.
- Conversation coaching. The same Virginia company uses Rilla, which records and analyzes technician conversations with customers so managers can coach for consistency.
- Call handling and lead follow-up. Tools like Netic handle inbound and outbound messages, book appointments automatically, and send maintenance reminders and slow-season text campaigns. Some contractors run chatbots after hours.
- Marketing. Contractors use ChatGPT, Canva and CapCut to turn out social posts and promotional material faster.
- Reporting. AI pulls numbers from several systems into one live dashboard of jobs, labor minutes and technician utilization.
Field service software has been moving the same way. At AHR Expo 2024, ACHR News reported on ServiceTitan features for dispatch, ad tracking, conversation coaching through a Rilla partnership, and alerts about missed leads so the office can call them back.
AI for HVAC technicians and field diagnostics
The most direct pitch to the person holding the gauges is the troubleshooting assistant. ACHR News described Bluon's MasterMechanic, a chatbot-style app trained on about 50,000 HVAC technical support calls, built to answer troubleshooting questions and help newer techs and quoting. Interplay Learning showed SAM, an AI advisor that walks trainees through problems with questions instead of handing them answers, built on Interplay's library of HVAC courses.
Treat the performance numbers with care. Bluon's creators said MasterMechanic answers troubleshooting questions with 96% accuracy and that the average tech saves about five hours a week. ServiceTitan tied its AI coaching to higher close rates and bigger tickets. All of those figures came from the companies selling the products, and the article didn't independently test them.
Contractors are more cautious. In the Plumbing & Mechanical piece, HVAC contractor Hugh Joyce said AI's effect on technicians' daily work is still limited. He expects it to become important for diagnostics, research and parts sourcing, but advised contractors to research vendors carefully and not overestimate what AI can do right now.
Some AI is already inside the equipment itself. ACHR News described a Mitsubishi Electric Trane sensor that scans a room for hot and cold spots and directs airflow, and a smart electrical panel that sets priorities between loads like EV charging and a heat pump. For techs, that means more diagnostics involve reading what a smart system decided and why, not just measuring what it's doing.
What this looks like by trade
| Trade | Documented uses | Still emerging |
|---|---|---|
| HVAC | Dispatch and routing, call handling, sales coaching, troubleshooting assistants, AI-driven comfort controls | Independent proof of diagnostic accuracy; parts sourcing |
| Plumbing | Office-side tools shared with HVAC shops: dispatch, booking, follow-up texts, marketing content | Field diagnostic tools specific to plumbing |
| Electrical | Office-side scheduling and marketing; smart panels that manage loads in homes | Code-question tools you can rely on without checking the code text |
The pattern holds across trades: AI is strongest where the task is repetitive, text-heavy or a scheduling puzzle. It's weakest where a mistake shows up as a failed inspection, a callback or a hazard.
Code lookup: where a chatbot shouldn't get the last word
Code questions are a natural thing to type into a chatbot. They are also where a confident wrong answer costs the most. The coverage we reviewed doesn't document any AI tool that contractors rely on for code compliance, and the Plumbing & Mechanical feature didn't address code accuracy or liability at all.
A safer workflow treats AI as a way to find the right section, not to settle the question:
- Ask the AI tool which section is likely to cover your situation.
- Open that section in the actual code text. NFPA, for example, offers NFPA LiNK for digital access to its codes and standards from any device.
- Confirm which edition your jurisdiction has adopted, and any local amendments. A model code is not necessarily the law where you're working.
- When the answer is unclear, ask the inspector or the authority having jurisdiction. Their call decides whether the job passes.
Rule of thumb: if you'd be uncomfortable telling an inspector "the chatbot said so," you haven't finished checking.
Estimating, proposals and paperwork
The sources we reviewed did not document contractors using AI for material takeoffs or detailed estimating. Where general-purpose chatbots fit well today is the writing around a job: turning rough notes into a clean proposal, drafting a maintenance-agreement explainer, writing a follow-up text, or rewording a scope of work so a homeowner understands it.
The same guardrails apply as with code. Quantities, prices and scope have to come from your own takeoff and your own price book. A chatbot can't see your supplier's pricing or the job conditions. Use it to write the words, and keep the numbers in your hands.
Low-risk ways to start this week
You don't need a software subscription to find out whether AI helps you. A general chatbot can handle a few low-stakes jobs right away:
- Clean up job notes. Paste in your rough notes from a service call and ask for a clear summary for the customer. Then read it and fix anything it got wrong.
- Explain a repair in plain English. Ask for a short explanation of why a part failed, written for a homeowner. You stay responsible for the facts.
- Study for a licensing exam. Ask it to quiz you on a topic, then check every answer against your code book or study guide. Wrong answers make good practice too, once you catch them.
- Draft routine messages. Appointment confirmations, maintenance reminders and follow-ups after a job are easy wins.
Leave customer names, addresses and account details out of anything you paste into a public chatbot unless you know how that tool stores your data.
Before you buy an AI tool: a checklist
Contractor and consultant Lynn Wise made a point in Plumbing & Mechanical that applies to every tool here: AI only works on clean, current data. She recommends cleaning up customer databases and price books before you start. A few more questions to ask any vendor:
- Who measured the results? Ask for numbers from customers like you, not company-wide averages from the vendor.
- What data is it trained on or connected to? A troubleshooting tool built on support calls for equipment you don't service won't help much.
- What does it cost per tech per month, and what's the cancellation term?
- Where does your customer data go? Know whether call recordings and customer details are stored or used for training.
- Who checks the output? Every AI message to a customer and every AI suggestion to a tech needs a human owner.
- Does it fit your size? Dispatch optimization pays off with many trucks; a two-person shop may get more from a good booking system.
What it means for your career
If you're early in a trade, the takeaway is not to learn AI instead of the trade. It's that the people who get the most from these tools already know the work well enough to spot a bad answer. Interplay's Ken Midgett pointed out to ACHR News that many manufacturer support staff just follow flowcharts without deep understanding. A tech who understands the system is the one who catches the AI's mistakes.
Practical skills worth building now: describing a problem clearly in writing, keeping tidy job notes, and checking a claim against a primary source. Those make you better with AI and without it. If you're moving toward a lead or office role, see the skills that matter on the way to lead.
Bottom line: AI is earning its keep in the office: dispatch, call handling, follow-up and marketing. In the field it's an assistant, best for training and narrowing a troubleshooting search, with accuracy claims that so far come mostly from vendors. Use it to work faster on words and scheduling, and verify anything touching code, safety or price against the real source.