← All guides Quotes and estimates

Rewrite a quote in plain English and answer "why so much?"

Turn a jargon-heavy quote into plain English and answer 'that seems high' calmly, without discounts. Tested on 4 AI tools.

Tested Oct 28, 2026 · 4 tools · Claude, ChatGPT, Copilot, Gemini

Trade: Cross-trade, HVAC, Handyman, Roofing
The job: “Make my quote easy for a homeowner to read, and answer a customer who says my price is too high”
Time saved: About 10 minutes per quote or reply
Tools tested: 4 · Claude, ChatGPT, Copilot, Gemini
Tested: Oct 28, 2026
Status

The job

Your quote makes sense to you. "R/R 3T HP split w/ matched AHU" means nothing to a homeowner. This guide does two jobs. First, rewrite the quote so the customer understands every line, with every price kept. Second, when a customer says "that seems really high," write a calm reply that explains the price without offering a discount.

Who it's for

HVAC, handyman and roofing owners whose quotes are full of trade shorthand.
Anyone who gets "I asked ChatGPT and it said $150" texts.
Owners competing against a cheaper bid.

The prompt

Use a new chat for each quote. Prompt 3 is optional and goes in the same chat as Prompt 2.

Prompt 1: plain-English rewrite

Rewrite this quote so a homeowner with no trade knowledge understands every line. Rules:
- Keep every line item, in the same order, with the exact same price. Do not merge, drop, or add lines.
- Short sentences. Explain any abbreviation or trade term in plain words.
- Do not add warranties, brand claims, timelines, efficiency savings, or code/permit statements that aren't already in the quote.

Quote:
[PASTE QUOTE]

Prompt 2: price pushback reply

A customer pushed back on my price. Write a calm reply text, under 120 words, that I can send. Explain what's included and why it costs what it does, using only the facts in my quote. Don't offer a discount. Don't badmouth anyone. Don't claim anything about the other contractor or about AI that you can't know. End with one question.

My quote:
[PASTE QUOTE]

Customer message:
[PASTE CUSTOMER TEXT]

Prompt 3 (optional): comparison checklist

Write a short checklist the customer can use to compare roofing quotes apples-to-apples, based on the items in my quote. Neutral tone, no sales pitch.

Paste your quote lines and prices where it says [PASTE QUOTE], and the customer's message where it says [PASTE CUSTOMER TEXT]. Prompt 3 was tested on a roofing quote. Change "roofing" to your trade.

Worked example

Input (fake)

Prompt 2, with this quote and customer text:

My quote:
1. Remove rotted lower door jamb & brickmould, left side (approx 18") — $165
2. Fabricate/install PVC jamb repair section, prime & caulk — $210
3. Adjust strike & weatherstrip — $60
4. Trip charge — $45
Total — $480

Customer message:
"That seems really high. I asked ChatGPT and it said this is a 30-minute job, maybe $150. Can you do better?"

Output

Claude · "Fable 5.1 High" · Max · tested 2026-10-28 · run 1, pushback reply on the door quote

Verbatim output
Hi Jake, fair question, happy to walk through it. The $480 covers four things: removing the rotted lower jamb and brickmould on the left side (about 18"), fabricating and installing a PVC repair section, priming and caulking it, and adjusting the strike and weatherstrip. $45 of that is the trip charge; the rest is labor and materials for the removal, the fabricated repair piece, and the finish work. It's a repair that has to be cut and fitted to your door, not an off-the-shelf part, which is where most of the cost comes from. Would it help if I broke down any of these line items further?

It's 107 words, offers no discount, and makes no claim about the customer's AI estimate. Check the line items against your quote before you send it.

Tool scores

A pass means usable as-is or after one small edit. Each tool ran 5 answers per run (two rewrites, two pushback replies, one checklist), 2 runs, so 10 answers each. Answers that didn't pass were scored partial.

Tool scores, from the guide frontmatter
Tool	Model (as shown)	Plan	Setting	Passed	With small edit	Note
Claude	“Fable 5.1 High”	Max	mode "Auto"	10/10Q1-Q7	0 of 10 passes	No misses. One soft promise: 'nothing on that list becomes an add-on later'.
ChatGPTThinking effort: High	not shown	not shown	Thinking effort: High	10/10Q1-Q7	0 of 10 passes	No misses. Once said the customer's ChatGPT estimate 'may not account for everything'.
Copilot	“Auto”	Personal	none	8/10Q1-Q7	0 of 8 passes	Both misses on the HVAC rewrite: '3-ton' and 'flush n/a' left unexplained or misread.
Gemini	“3.6 Flash”	not shown (Upgrade button visible)	none	6/10Q1-Q7	1 of 6 passes	Durability promises in 3 pushback replies ('so the rot doesn't return'); jargon left once.

ChatGPT doesn't show which model answers; tested on the High setting.

The main finding: every tool kept every line and every price in every rewrite, and none offered a discount. The risk is what gets added. Gemini promised results the quote never did ("so the rot doesn't return," "long-term durability"). Copilot left HVAC jargon unexplained and misread "flush n/a." Claude and ChatGPT had only small notes, like one soft promise and one mention of the customer's AI estimate.

What you still check

Go down this list before anything reaches the customer.

Every line, in the same order, with the same price. Count the lines and add up the total yourself.
Each line still means the same thing. Renames like "trip charge" to "service call fee" are fine only if you'd say it that way.
Every trade term is explained. Watch equipment lines: "3-ton," "15.2 SEER2," "flush n/a."
No promises about how long the work lasts ("so the rot doesn't return," "sealed long-term"). Strike them unless your quote says so.
No soft promises you didn't make ("nothing on that list becomes an add-on later").
Nothing about the customer's AI estimate or the other contractor beyond what the customer told you.
The reply says why it costs that much, not just a list of items. Use the real cost drivers from your quote, like the trip charge or a custom-fit part.
One reply, ready to send: no "[Customer Name]" placeholders, and pick one if the tool gives you two options.
Comparison checklists can add items (warranty, decking). Be ready to answer every item on it.
Claude may put the reply in an editable draft card. Copy it from the card. Gemini may offer to email the checklist to the customer. Send it yourself, from your own phone or email.

Limits

Tested on three fake quotes only (HVAC replacement, door repair, reroof vs. a cheaper bid). Your trade's shorthand may trip the tools in new ways.
Results vary from run to run, by plan and setting, and as the apps update. These scores are from 2026-10-28.
Two runs is a small test. A tool that passed here can still miss on your next try.
Gemini's run 2 checklist answer came from its "Thinking" model, not "3.6 Flash". The picker switched on its own.
Nothing here is legal, permit, warranty or building-code advice. If your quote has a permit line, the AI should explain it, not add to it.
The AI output is never guaranteed. You are the one sending it.

How we tested

Date: 2026-10-28, about 10:15 to 10:55 AM CT.
Route: each tool's website, signed in, a new chat for each quote.
Tools, model labels and plans (as the apps showed them):
- Claude: "Fable 5.1 High", mode "Auto", plan Max.
- Copilot: "Auto", plan Personal.
- ChatGPT: Thinking effort: High. No plan label shown. ChatGPT doesn't show which model answers; tested on the High setting.
- Gemini: "3.6 Flash". Plan not shown (Upgrade button visible). Run 2's checklist answer came from "Thinking"; the picker switched on its own.
Runs: Prompt 1 on the HVAC and door quotes, Prompt 2 on the door and roof quotes, Prompt 3 on the roof quote. Two runs, so 10 scored answers per tool.

A pass means usable as-is or after one small edit.

Pass criteria. Each answer was scored on the criteria that apply to its prompt:
- Q1: the rewrite keeps every line and price unchanged.
- Q2: readable, with jargon explained (R/R, 3T, SEER2, AHU, kW, lineset, whip, brickmould).
- Q3: no new claims (warranty, efficiency savings, brand, code, timeline).
- Q4: the pushback reply is 120 words or less, calm, with no discount.
- Q5: cost drivers named only if they're true per the quote.
- Q6: no badmouthing the competitor; their items described only as the customer described them.
- Q7: the comparison checklist is neutral and covers underlayment, ice & water, drip edge, boots, ventilation and haul-off.
Checking: a script checked prices, line counts, word counts and trigger words, then a person read every answer.
Fake data only. All test data was made up. No real customers, jobs or prices were used.

The digest

One job, done with AI, every week.

A short email with one tested guide and what to double-check. Free. Unsubscribe anytime.

Subscribe freeOpens our beehiiv signup page. We don't sell your email.

Found something wrong? Send a correction.