82% of HVAC businesses reported using AI in some part of their operation in 2026; the most common uses were estimates and quotes (58%) and invoices and billing (57%).Source ↗
In a 2026 survey of 1,032 trade contractors, only 12% had fully embedded AI into their workflows while 34% were still actively experimenting — adoption is real but early.Source ↗
Among contractors already using AI, 62% reported measurable gains in efficiency and productivity, with many saving three or more hours per week.Source ↗
Handling refrigerant legally requires EPA Section 608 technician certification under the Clean Air Act — a licensing responsibility no AI tool can stand in for.Source ↗
writingClaudeChatGPT

Turning a scope of work into an itemized estimate customers understand

Estimates and quotes are the single most common thing HVAC shops point AI at (58% of businesses that use it). After a diagnosis or a walk-through you know the scope and your prices, but writing it up so a homeowner sees what they're paying for eats the evening. The catch every guide repeats: AI will happily invent prices and specs, so it drafts the wording while every number stays yours.

Prompt
You are a proposal writer for {{company_name}}, a licensed HVAC contractor. Turn my job notes into a clear, itemized estimate a homeowner can read and approve.

Job scope (what I diagnosed and what I'm proposing): {{job_scope}}
Line items with MY prices (labor, parts, equipment): {{line_items}}

Write the estimate with:
1. A one-paragraph plain-language summary of the problem and the recommended work.
2. An itemized table: description, quantity, price — using ONLY the line items and prices I gave you.
3. A short "what this includes / does not include" note.
4. A neutral closing line inviting questions and explaining how to approve.

Hard rules:
- Use ONLY the prices, quantities, model numbers, and tonnage I provided. Never invent or estimate a price, part number, SEER rating, capacity, or code requirement. If something needed for the estimate is missing, insert [ADD: what's missing] instead of guessing.
- Do not add warranty terms, rebates, financing offers, efficiency claims, or guarantees I did not state.
- Plain language, no pressure tactics, no "act now." Under 250 words plus the table.

Fill in your details and the prompt updates live — then copy.

communicationClaudeChatGPT

Explaining repair-versus-replace so a homeowner can actually decide

The hardest conversation in residential HVAC isn't technical — it's the kitchen-table moment where a homeowner has to choose between a $2,400 repair on an aging system and a $7,000 replacement, and they suspect you're upselling. Terms like static pressure, refrigerant phase-out, and system sizing mean nothing to someone who just wants the house cool again. AI is good at translating your findings into a fair, plain-language rundown — as long as it explains, and you decide.

Prompt
You are helping a licensed HVAC technician write a clear, honest message to a homeowner explaining their repair-versus-replace options. Keep it balanced — the goal is an informed customer, not a hard sell.

What I found (my diagnosis): {{diagnosis}}
The options with MY prices: {{options}}
About the customer: {{customer_context}}

Write a message that:
- Explains the problem in plain language, translating any technical term in a few words.
- Lays out each option I gave — what it costs, what it fixes, and roughly how long it buys them — using ONLY my prices and facts.
- Includes a neutral "what happens if you do nothing" line based on what I described.
- States my recommendation ONLY if I included one; otherwise presents the options evenly and says the choice is theirs.

Hard rules:
- Use only the findings and prices I provided. Do not invent efficiency percentages, energy-bill savings, rebate amounts, equipment lifespans, or refrigerant facts. If a number would strengthen the message, write [VERIFY: what] instead of inventing it.
- No pressure language, no scare tactics. Warm, straightforward, under 180 words.

Fill in your details and the prompt updates live — then copy.

creativeChatGPTClaudeGemini

Marketing maintenance plans that fill the slow season

Maintenance agreements are what smooth out a shop's cash flow between the summer and winter rushes, but the emails and social posts that sell them are the first thing that gets skipped after a three-call day. AI drafts the marketing fast; the risk is that it reaches for hype and invents guarantees — exactly the claims that get a contractor in trouble. Give it your real plan terms and hold it to accurate, plain language.

Prompt
You are a marketing writer for {{business_name}}, an HVAC company serving {{service_area}}. Write content promoting our maintenance plan for the {{season}} season.

Our plan — use ONLY these facts: {{plan_details}}

Produce:
1. One short email (under 150 words) to existing customers explaining the plan's value and how to sign up.
2. Three social posts (under 60 words each): one practical seasonal tip, one that names a concrete plan benefit, one friendly reminder.
3. Two subject-line options for the email.

Hard rules:
- Use only the plan features, prices, and terms I listed. Do not invent benefits, discounts, visit frequencies, or coverage.
- No guarantees about equipment lifespan, energy savings, or breakdown prevention. If a statistic would help, write [STAT: verify] rather than inventing one — maintenance reduces the odds of failure, it doesn't promise anything.
- Plain, friendly language. Banned words: revolutionize, game-changer, unleash, supercharge, transform. No false urgency or fake scarcity.

Fill in your details and the prompt updates live — then copy.

automationChatGPTClaude

Review requests, on-my-way texts, and seasonal reminders you write once

Most of the customer messages a shop should send never go out, because the person doing the install at 3 p.m. is doing invoices at 9 p.m. Housecall Pro and other field-service tools let AI draft the repeatable templates — on-the-way texts, review asks, seasonal reminders — that you load once with merge fields and let your software send. Write them well one time and stop reinventing them per customer.

Prompt
You are writing reusable customer-message templates for {{company_name}}, an HVAC company in {{service_area}}. These load into field-service software with merge fields — use [CUSTOMER_NAME], [TECH_NAME], and [ARRIVAL_WINDOW] as the only placeholders. Do not invent any other merge fields or any facts about the company.

Write this set:
1. "On my way" text: friendly, says who's coming and the arrival window, under 30 words.
2. Job-complete text: thanks them, says the work is done, and asks for a review on {{review_platform}} only if they were happy — one link, no pressure.
3. Seasonal maintenance reminder (email, under 90 words): reminds them it's time for a {{season}} check-up and how to book.
4. Missed-appointment / reschedule text: warm, no guilt, offers the next slot.

Rules: plain and human, no corporate stiffness, no "just checking in," no false urgency, no discounts unless I add them. Nothing that guarantees equipment condition or implies the last visit fixed every possible issue.

Fill in your details and the prompt updates live — then copy.

analysisClaudeChatGPT

Finding the patterns hiding in your reviews and callbacks

A shop owner reads reviews one at a time and reacts to whichever one stung most this week, so the same recurring complaint — techs running late, confusing invoices, a specific install that keeps generating callbacks — never gets named as a pattern. AI is genuinely useful at grouping a pile of feedback into themes. It's a summarizer here, not a judge: it tells you what people said, and you decide what to fix.

Prompt
You are analyzing customer feedback for the owner of an HVAC company. I'll paste a batch of reviews, survey replies, and callback notes. Summarize the themes — do not evaluate whether we're a good company or invent explanations.

Feedback (names and addresses removed): {{feedback_batch}}
Time period: {{time_period}}

Produce:
1. The top 3-5 recurring themes, each with how many mentions and one representative (paraphrased) quote.
2. A split of what customers praised versus what frustrated them.
3. Anything that looks like a repeat operational issue (scheduling, communication, pricing clarity, callback on a specific job type).
4. Two or three concrete questions I should ask my team based on what's here.

Hard rules:
- Summarize ONLY what's in the text I pasted. Do not invent complaints, causes, or fixes, and do not guess at a diagnosis for any technical issue mentioned.
- If there isn't enough feedback to call something a pattern, say so instead of stretching.
- Neutral and factual — you're organizing what customers said, not grading my business.

Fill in your details and the prompt updates live — then copy.

planningClaudeChatGPTGemini

Prepping for a service call with a model-specific briefing

A newer tech heading to an unfamiliar rooftop unit or a discontinued furnace brand often spends the night before digging through forums and manuals. AI can pull together a quick briefing — common failure points, where to look, what to ask the customer — that gets you oriented faster. The hard line: this is background reading, not a diagnosis. It knows nothing about the actual unit, and refrigerant, sizing, and code calls stay on site with you.

Prompt
You are a research assistant helping a licensed HVAC technician prepare for a service call. Build a pre-visit briefing. You know NOTHING about this specific unit or home — everything you produce is general background to verify on site.

Equipment: {{equipment_make_model}}
System age / context: {{system_age}}
Reported symptom from the customer: {{reported_symptom}}

Produce:
1. Common, well-documented failure points for this type/era of equipment — framed as "commonly seen on units like this, verify on site."
2. What to inspect and where to look for each, and what to confirm before concluding anything.
3. Two or three questions to ask the customer that would narrow things down.
4. A "don't assume" list of things techs commonly misattribute on equipment like this.

Hard rules:
- Never state that THIS unit has any specific fault, refrigerant type, or cause. Everything is "commonly," "often," "verify."
- Do not provide refrigerant charge amounts, wiring specifics, or code requirements as fact — direct me to the unit's data plate and the manufacturer's service literature for those.
- Flag anything safety-critical (combustion, electrical, refrigerant) as "confirm with proper testing and PPE." No diagnosis, no repair steps I should follow blind.

Fill in your details and the prompt updates live — then copy.

Common questions from hvac technicians

Is it OK for HVAC technicians to use AI on the job?

For the business and customer side — estimates, proposals, customer messages, marketing, review responses — it's widely used and fine, and 82% of HVAC businesses report using AI somewhere in their operation. Keep it away from the licensed side. Refrigerant handling (EPA Section 608), load calculations, equipment sizing, and code compliance are your professional judgment and carry your liability; a chatbot's output doesn't change who signs off.

Can AI size a system or run a Manual J load calculation?

No — not reliably, and not in a way you can put your name on. Manual J and equipment selection depend on real measurements of the specific home and are your licensed responsibility. A language model works from general text and will hand you confident, wrong tonnage, refrigerant, and code answers. Use purpose-built load-calc software and your own field data, not a general AI tool.

Can I paste customer names, addresses, or payment info into ChatGPT?

Don't. Consumer AI tools may retain what you enter, and customer contact and payment details don't belong there. Use placeholders or merge fields (like [CUSTOMER_NAME]) in your prompts and add the real details only inside your own field-service software, ideally one with a clear no-training data policy.

A customer says ChatGPT already diagnosed their system — how do I handle it?

Politely, and with your gauges. AI can sound completely certain and still be wrong — techs report customers arriving convinced "it's the capacitor" when it wasn't. Acknowledge what they read, then explain that a real diagnosis comes from testing the actual unit. Your EPA certification and license mean your measured findings, not a chatbot's guess, are what the repair is based on.

Related professions