ChatGPT prompts for contractors and home service businesses: 18 copy-paste prompts from real roofers, handymen, landscapers, electricians and cleaners (2026)

What ChatGPT prompts help contractors with estimates and proposals?

Four businesses use it around the quote. Craig Timm runs each finished estimate through ChatGPT; it points out what he forgets, like markup or a left-out part, and on the estimates he has tested it always adds about 30% to certain areas, so he has started marking up materials. J.R. Crowell's team wears a Plaud recording pin on remodel walkthroughs and feeds the dictation to a chatbot he built around the International Residential Code, which builds the scope and the material list. Monarch Landscape Management had ChatGPT write the benefits of services like turf aeration once, as a template for every proposal. Emily May photographs the customer's house and asks ChatGPT which color in the shingle line she installs fits best, then turns her phone around to show them.

Check this handyman estimate before I send it. Job: {SHORT DESCRIPTION}. Line items: {LINE ITEMS WITH HOURS, MATERIALS AND PRICES}. My rules are in the project instructions. List: 1) parts, supplies or steps that are probably missing for this job, 2) where my labor hours look low for the work described, 3) whether markup, trip fees or disposal are left out, 4) anything a customer might misunderstand. Show the math for any change and don't change my hourly rate.

This is the transcript of a remodel walkthrough at {ADDRESS OR JOB NAME} for {COMPANY NAME} in {CITY/COUNTY}, where {CODE EDITION, e.g. the 2021 IRC with local amendments} applies. Using my scope template: 1) list every measurement I called out, by room, 2) write the scope of work in order, 3) give a material takeoff with quantities (framing, hangers, fasteners, drywall, trim), 4) flag anything that needs to change to meet code, with the section I should check, and 5) list questions to ask the homeowner. Mark anything you inferred rather than heard as [CHECK].

Write a standard proposal section for {SERVICE, e.g. core aeration} from {COMPANY NAME}, a commercial landscape maintenance company in {CITY}, for property managers. Use these facts only: {WHAT IT IS, HOW WE DO IT, TIMING, WHAT THEY'LL SEE AFTERWARD}. Structure: What it is (2 sentences), How it works (3 bullet points), What to expect (before, during, after) and Why now (1 sentence). Plain language, no hype, no claims I didn't give you.

Here is a photo of a customer's house. We install {SHINGLE LINE, e.g. Atlas Pinnacle Pristine}. Which three colors in that line would look best with this house's brick or siding, trim and front door, and why, in one sentence each? Then show the house with the roof changed to your top pick, keeping everything else in the photo the same.

What ChatGPT prompts help with job costing, budgets and invoices?

Cory Overman uploaded a season of production data from his CRM, and ChatGPT worked out the average cost of a single and a double aeration pass and the seed needed per square foot. Cory Shorette set up a paid ChatGPT project told "You are my CFO", loaded with the company's financials, which helped build the 2026 budget; he says owners usually know the right decision, "but it's nice to have a double-check and backup of your thesis." Morgan Sterling used ChatGPT to help build a script that pulls subcontractor invoices from email, reads them, identifies the job and files them, and reports saving about 5–6 hours a week.

Attached is a season of {SERVICE, e.g. aeration and overseeding} jobs from {COMPANY NAME} (no customer names): lawn size in square feet, crew hours, number of passes, seed and material used, and price charged. 1) Work out average labor and material cost per 1,000 sq ft for a single pass and for a double pass. 2) Work out seed used per 1,000 sq ft. 3) Flag rows that look like data-entry mistakes. Show the formulas you used so I can check them in Excel. Don't suggest a price; I'll add overhead and margin myself.

You are the CFO of {COMPANY NAME}, a {TYPE OF LANDSCAPING / SERVICE} company in {CITY} with about {REVENUE} revenue and {NUMBER} crews. Work only from the uploaded financials. For every answer: 1) show the calculation in a table, 2) pressure-test it by checking it against the uploaded actuals, 3) list every assumption you made and what number would confirm it, 4) say plainly if the data isn't enough. Never invent a figure. Task: {TASK, e.g. BUILD NEXT YEAR'S BUDGET WITH OVERHEAD LINE ITEMS}

Read this subcontractor invoice and return JSON: vendor, invoice_number, invoice_date, due_date, amount_total, job_name_or_address, description (one line), confidence (high/medium/low).
Match the job to one of these active jobs: {JOB LIST}. If no match, set job_name_or_address to 'UNMATCHED'.
Invoice text: """{INVOICE TEXT}"""

What ChatGPT prompts help in the field?

Craig Timm photographed loose thermostat wires and ChatGPT gave him the model number and wiring schematic, asked for a photo of the heating unit, then told him where every wire went; he uses the same photo search for faucet parts and parts for older houses. Toby Mitchell replaced his safety software with ChatGPT-written toolbox talks about what crews actually face, such as driving in the fog, after the software kept suggesting topics like hypothermia in a California July.

I'm a handyman on a job. Attached are photos of {WHAT IT IS, e.g. a bathroom faucet whose handle won't come off / thermostat wires that were disconnected}. Tell me: 1) the likely brand and model and how sure you are, 2) the part I probably need with its part number, 3) step-by-step how to do the repair, with any safety step first, and 4) what photo or label you'd need to be more certain. If you're guessing, say so.

Write a five-minute toolbox talk for an electrical crew at {COMPANY NAME} in {AREA}. This week: {WHAT'S HAPPENING, e.g. fog on the morning drive, attic work in the heat, a near-miss on a ladder, a new apprentice}. Give 4 to 6 short talking points in plain language tied to that situation, one real-world example, and 3 questions to ask the crew. Don't cite specific OSHA or NEC section numbers; I'll add those myself after checking them.

What ChatGPT prompts help write SOPs, training programs and manuals?

Scott Parker talks his ideas out to ChatGPT and has it break them into SOPs, and had ChatGPT and Claude evaluate the employee handbook an HR company helped him write. Derek Taussig gave ChatGPT his state's rules and his training manuals and built the chemical-application training program he had put off for years. Lachlan Walker built a custom GPT on ChatGPT Plus with his operations manuals and contracts; it writes each new site's operations manual.

1) Turn what I just said into a standard operating procedure for {TASK, e.g. a residential service call} at {COMPANY NAME}: purpose, who does it, tools and materials, numbered steps, safety checks, and what 'done' looks like. Use only what I said and mark anything I left out as [ASK]. 2) For the attached employee handbook: list topics a small electrical contractor's handbook usually covers that this one is missing or unclear on (with the page), and anything that contradicts itself. Don't rewrite it and don't give legal conclusions.

Attached are {STATE}'s rules for {CATEGORY, e.g. commercial ornamental and turf} pesticide applicator training, our product labels and our current training manual at {COMPANY NAME}. Build a training program outline that covers every requirement in the rules: modules, learning goals, hands-on checks and a short quiz for each module. Cite the rule section each module covers. Use only what's in the attached documents, and list any requirement you couldn't find covered.

Using our master operations manual and the example site manual in this project, write the operations manual for a new site for {COMPANY NAME}. Site details: {SITE TYPE, SIZE, AREAS, CLEANING FREQUENCY, SHIFT TIMES, SPECIAL REQUIREMENTS}. Include: scope by area and frequency, shift plan, supplies, safety and access procedures (leave codes as {CODE}), quality checks and who to call. Follow our manual's format; mark anything the site details don't tell you as [ASK SUPERVISOR].

What ChatGPT prompts help with customer messages and reviews?

Cody Jackson writes his own message and asks ChatGPT to soften it when it sounds too direct. Dallas Sunrise Maids uses ChatGPT for client complaints, schedule change requests and basic emails, which helps office staff whose first language isn't English. Shades of Green's technicians send pre-written texts, which ChatGPT helps write, when they spot a lawn problem like dollar spot. Centennial Roofing drafts Google review replies with ChatGPT and checks them against Gemini's.

Here's a message I wrote to {WHO, e.g. a homeowner / a crew lead / a supplier}: {MESSAGE}. I'm a direct person and I don't want this to come across as harsh. Make it a little softer and friendlier, keep it about the same length, keep every fact, date and price exactly as written, and don't add anything I didn't say. Then list what you changed.

Draft a reply from {COMPANY NAME}, a house cleaning company, to this client email: {EMAIL}. Our policies: {RE-CLEAN POLICY, CANCELLATION WINDOW, DAMAGE POLICY}. Our tone: {TONE, e.g. warm, short, no corporate words}. If it's a complaint, apologize once, say exactly what we'll do and when, and don't offer anything outside our policies. If it's a schedule change, confirm the new date only if it's in this list of open slots: {OPEN SLOTS}. Under 120 words. Also give me the same reply in {STAFF LANGUAGE} so I can check it.

Write a text message from {COMPANY NAME}, a lawn care company in {AREA}, to a customer whose lawn our technician just found has {CONDITION, e.g. dollar spot}. In under 320 characters: say what we saw in plain words, why it happens this time of year, what the customer can do (watering, mowing height), and that we offer {SERVICE} if they'd like help, with no pressure. Friendly, not salesy. Don't promise results.

Draft a reply from {COMPANY NAME}, a roofing company in {CITY}, to this Google review: {REVIEW TEXT}. Thank the customer by first name, mention one specific thing they said, and keep it under 80 words in a warm, plain tone. If the review is negative, apologize once, don't argue, and invite them to call {OWNER OR MANAGER NAME} at {PHONE}. Don't mention insurance claims, prices or anything they didn't write.

What ChatGPT prompts work for contractor marketing?

Benjamin Schneider asks ChatGPT for a Facebook post about one service, posts it with a photo of his work on his page and in local groups, and spends no more than an hour a week on marketing. Rebecca Stubbs gets newsletter topics and a basic framework from ChatGPT and rewrites them for Jonny Nichols Landscape Maintenance's customers.

Write a Facebook post for {BUSINESS NAME}, a handyman business in {TOWN}, promoting {SERVICE, e.g. furniture assembly}. It goes with a photo of {WHAT THE PHOTO SHOWS}. Two short paragraphs, friendly and local, a few emojis, 3 hashtags, and end with how to book: {PHONE OR WEBSITE}. Don't invent prices, guarantees or reviews.

I write a monthly email newsletter for customers of {COMPANY NAME}, a lawn and landscape company in {TOWN AND STATE}. Next month is {MONTH}; our crews will be doing {SERVICES}. Recent jobs: {ONE OR TWO REAL JOBS}. Suggest 8 newsletter topics homeowners here would actually care about, each with a one-line hook. Then outline the top 2 in 4 or 5 bullet points each. No made-up statistics or product claims.

What should contractors watch out for with ChatGPT?

• It can be wrong. On the podcast, Craig Timm and the host agree AI sometimes gives wrong information, so he researches further rather than taking it as the final answer (case). Kill the power and follow the manufacturer's instructions before touching wiring.
• Code, permits and state rules. A scope or training program built from the code or state rules still needs your own check; Taussig says his program is state-approved, and yours needs the same sign-off before it counts (case).
• Prices. Use your own prices and costs; a chatbot has no live price data for your town. Timm treats its estimate check as a trial, not a final answer.
• Money decisions. Lachlan Walker rated its finance help "Decent" but says his bookkeeper does far better and he won't base important decisions on it (case).
• Your voice. Walker rated its marketing "Bad" because it reads as AI-written, Schneider warns against letting AI write a whole website, and Jackson says owners who let AI write their contracts and emails lose the personal touch.
• Customer data. As a suggestion, remove customer names, addresses and gate or alarm codes before you upload anything, as the Shades of Green and Walkers prompts do.