You finish the site visit. You've got the job scoped in your head: what it'll take, what it'll cost, how long the crew needs. But the quote? That goes on the list. You'll write it up tonight. Or maybe tomorrow. By the time it lands in the customer's inbox, two days have passed — and sometimes the job has already gone to someone who responded faster.
Slow proposals cost real money. A ServiceTitan survey of 1,000 residential contractors published in April 2026 found that 74% see AI as key to efficiency. In ServiceTitan's broader 2026 State of AI in the Trades report, only 12% of contractors had actually embedded AI into their day-to-day operations. The gap between knowing and doing is exactly where contractors are losing time they can't get back.
Writing estimates is one of the places AI can help the most, starting today, without special software or a tech team. Here's how.
What AI is actually doing here
AI isn't replacing your pricing judgment or your knowledge of the job. It's doing the writing — the translation from rough job notes into a professional, readable document that looks like you spent 45 minutes on it.
The gap AI fills is the drafting step. Most contractors know exactly what goes into a quote; they just don't want to stare at a blank page after a long day on site. A solid AI prompt turns three sentences of rough notes into a clean scope of work, a line-item summary, and a professional cover paragraph in under two minutes. You review it, adjust the numbers, and send it while the visit is still fresh.
According to that State of AI in the Trades report, 62% of contractors already using AI report measurable gains in efficiency, with many saving three or more hours per week. Proposals are one of the biggest recurring time sinks that AI consistently helps trim.
A four-step process that works on any phone
- Record a voice note immediately after the site visit. While you're still in the driveway, describe the job out loud: what you saw, what you'll do, rough material quantities, any access issues, timeline, and assumptions. Don't worry about structure — talk like you're leaving a message for your office manager.
- Transcribe it. On iPhone: Voice Memos → Share → Transcribe. On Android: Google Recorder handles transcription automatically. Either way, this takes under a minute.
- Paste the transcript into Claude or ChatGPT with a prompt. (More on this below.)
- Review, adjust pricing, and send. The AI draft is a starting point, not a finished document. Check every line item against your actual numbers, add your payment terms and exclusions, and send it. Total time: 5–10 minutes instead of 45.
Writing a prompt that gets useful output
The quality of your draft depends on the quality of your prompt. A bare-bones instruction like "turn this into a quote" produces a bare-bones result. A specific prompt produces something close to sendable.
Here's a starting template you can adjust for your trade:
"You are writing a professional estimate for a small [plumbing / electrical / HVAC / landscaping] company. The customer is [brief description, e.g. 'a homeowner in a 1970s bungalow']. From the notes below, write: (1) a one-paragraph scope summary in plain English, (2) a line-item list of the main work items, and (3) a short closing paragraph that mentions the timeline and next steps. Do not include prices — I will add those. Use a friendly, professional tone. Avoid jargon the customer wouldn't understand. Notes: [paste your transcript]"
Once you have a version that consistently produces good output for your type of work, save it as a text snippet you can paste in every time. The prompt is where your experience lives — be specific about your trade, your tone, and what you always include or always exclude.
If you want a tool that goes further — combining AI-assisted content with e-signature, payment collection, and a client-facing view — PandaDoc starts at $19/user/month (billed annually; $35 month-to-month) and handles the full proposal-to-signed-contract workflow. Proposify is a similar option with stronger analytics, starting at $19/user/month billed annually ($29/user/month month-to-month). Both add more process than most small operators need — but once you're sending more than 10 proposals a week, the template management and tracking features start earning their keep.
The three things to always check before you send
AI drafts the words. You own the numbers. Never send an AI-generated proposal without checking these:
- Pricing accuracy. The AI has no idea what materials cost this week or what your labour rate is. Fill in every number yourself — do not let an AI guess at a price and leave it in.
- Scope completeness. Re-read the draft against your notes. AI occasionally skips a line item or combines two things that should be separate. Compare it to your voice note before you add prices.
- Exclusions, stated clearly. Add them explicitly. What you are not doing is often as important as what you are — particularly for disputes down the road. A short exclusion list also signals professionalism to the customer.
The right mental model is this: the AI draft is what a capable office assistant would hand you to review. You'd still read it before it went out, and so should you here.
What to do this week
- After your next site visit, record a two-minute voice note in the driveway — full scope description, unfiltered.
- Transcribe it and paste it into ChatGPT or Claude with the prompt template above, adjusted for your trade and tone.
- Add your pricing and exclusions, review once, and send.
- Note how long that took compared to your usual process.
The contractors gaining ground right now aren't always the most skilled at the work — they're the ones responding the fastest with a clear, professional proposal. The ServiceTitan data shows the majority of contractors see AI as important to their business, but most haven't started. That gap is the opportunity. You already have the knowledge. AI just helps you get it out of your head and into the customer's inbox before someone else does.