How to use AI to turn your existing price list into a quote or estimate you can send to a client in under five minutes

Manual quoting eats 20–60 minutes per job for most service businesses — time that Salesforce research{:target="_blank"} confirms is already scarce, given that sales reps spend only 28% of their week on actual selling. This post walks you through a repeatable, AI-assisted system for how to use AI to write a quote fast — converting your existing rate card into a professional, client-ready estimate using ChatGPT or Claude — no new software required. Do it right once, and every quote after that takes under five minutes and arrives in the client's inbox before your competitor has finished opening a spreadsheet.

What you need before you start

ChatGPT{:target="_blank"} — OpenAI's conversational AI, which supports direct file uploads of PDFs, spreadsheets, and Word documents so you can feed it your existing price list and output a formatted quote. Pricing: The free tier (GPT-3.5) works for basic text generation, but you need ChatGPT Plus{:target="_blank"} at $20/month (as of April 2025) to access GPT-4o and its file upload capability. If you prefer an alternative, Claude{:target="_blank"} by Anthropic — specifically the Sonnet 3.5 or 3.7 models — handles structured document output equally well and is available on the Claude Pro{:target="_blank"} plan at $20/month (as of April 2025). Both free tiers allow text-paste inputs, which covers most use cases if your price list is short enough to copy.

Time required: 15–20 minutes for initial setup (building your master prompt and testing it once); under 5 minutes per quote after that.

Skill level: No technical background needed. You need to be able to copy and paste text, and optionally upload a file. If you store your rate card in Google Docs or Sheets, Google Gemini{:target="_blank"} — integrated directly into Google Workspace — is also worth considering, particularly if you're already on a Google Workspace Business plan (pricing starts at $14/user/month as of April 2025).

How to use AI to write a quote fast from your existing rate card

1. Open ChatGPT (GPT-4o) or Claude in your browser and start a new conversation.

2. Upload or paste your price list. If it's a PDF or spreadsheet, use the file upload button. If it's a short text list, paste it directly into the chat. Label it clearly at the top of your message: "Here is my rate card. I'll follow with a quoting request."

3. Write your quoting request directly below the rate card using the master prompt format in the next section. Include the client name, the services they've asked for, quantities or estimated hours, any discount, and your preferred tone.

4. Review the output before you copy anything. Check that every line item matches your actual rate card, that the subtotal arithmetic is correct, and that the scope description accurately reflects the job. AI models can misread a figure or conflate two similar line items — verify the numbers against your source.

5. Copy the output into your Google Docs or Word quote template. This step takes under two minutes if your template already exists. Drop in your logo, adjust any formatting, and add tax manually (see the note below on tax).

6. Save the final document as a PDF and send it — via email, a client portal, or a tool like DocSend{:target="_blank"} if you want read-receipt tracking.

On tax: ChatGPT and Claude do not apply jurisdiction-specific sales tax rules automatically. Add your applicable tax rate as a manual line item after generation. Do not rely on the AI to calculate this correctly.

Skipping step 4 — the review — is the single most common mistake in AI-assisted quoting. The output looks professional, which creates false confidence. A misread hourly rate or a missed line item that slips through to a client damages your credibility more than a slow quote does.

How to write the master prompt that does the heavy lifting every time

The master prompt is a reusable template you save — in a notes app, a Google Doc, or directly in a tool like Notion{:target="_blank"} — and paste at the start of every quoting session. It encodes your business name, your standard terms, and your preferred output structure so you don't have to re-explain them each time.

Here's a working template:

You are a quoting assistant for [Your Business Name].

Using the rate card I have provided, create a professional client estimate with the following details:

• Client name: [Client name]
• Project or job description: [Brief description, e.g., "bathroom tile installation, approximately 40 sq ft"]
• Services requested: [List each service by name as it appears on the rate card]
• Quantities or hours: [E.g., "6 labour hours, 40 sq ft of tile supply"]
• Discount (if any): [E.g., "10% loyalty discount on labour"]
• Tone: [Formal / Professional-friendly]

Output format:

1. A brief introductory paragraph addressed to the client (2–3 sentences, professional tone)
2. An itemised table with columns: Description | Qty/Hours | Unit Rate | Line Total
3. A subtotal row
4. A placeholder row for tax: "Tax (to be confirmed): $___"
5. A total row
6. A paragraph noting that this estimate is valid for 30 days and outlining your standard payment terms: [Insert your payment terms here]
7. A clear call to action: [E.g., "Please reply to confirm you'd like to proceed and we'll schedule the work."]

Do not invent rates. Use only the figures from the rate card I have provided.

The last instruction — "do not invent rates" — matters. Without it, some models will interpolate a plausible-sounding figure when a service is ambiguously named on your rate card. Flagging the constraint reduces that risk significantly.

When something goes wrong

The AI invents a line item or uses the wrong rate. Root cause: The service name in your request doesn't exactly match the name on your rate card, so the model makes a reasonable but wrong assumption. Fix: Standardise your service names on the rate card and use those exact names in every prompt. If your rate card says "Interior latex paint — per room," don't request "painting — bedroom." Match the label precisely.

The subtotal doesn't add up correctly. Root cause: Multi-line arithmetic errors are rare but not impossible, particularly on longer quotes. Fix: Paste the itemised table into a Google Sheet and use a SUM formula to independently verify the total. This takes 45 seconds and removes any residual doubt before you send.

The formatting breaks when you paste into Word or Google Docs. Root cause: AI-generated Markdown tables don't always render cleanly in word processors. Fix: Ask the model to output the table in plain text with tab-separated columns, or copy into Google Sheets first and then paste special (values only) into your document. Alternatively, Claude's output tends to produce cleaner copy-paste results than ChatGPT for structured tables — worth testing if this is a recurring issue.

What to do next

Build your master prompt, run it on one real job, and compare the result against your last manually produced quote — time the process both ways. That's the only benchmark that matters for your business. Once the prompt is dialled in, the next logical step is connecting your quoting workflow to your client follow-up process, so accepted quotes automatically trigger an onboarding task or deposit request.

For that kind of automation, see our guide on connecting AI tools to your client onboarding workflow.

If you're producing more than 15–20 quotes a month and need e-signature, CRM integration, and a proper audit trail, look at PandaDoc{:target="_blank"} — their Essentials plan starts at $19/user/month (as of April 2025, check current pricing{:target="_blank"}) and adds tax logic, payment collection, and document tracking that pure prompt-based quoting can't match.

FAQ

Can I use the free version of ChatGPT to generate quotes? Yes, with limitations. The free tier runs GPT-3.5, which handles text-paste prompts competently but doesn't support file uploads. If your rate card is short enough to paste as text — under roughly 2,000 words — the free tier is sufficient for the method described here. For file uploads and more reliable formatting on longer documents, you need ChatGPT Plus at $20/month (as of April 2025).

How accurate are AI-generated quote totals? Accurate enough to use as a starting point, not accurate enough to send without review. Arithmetic errors on simple itemised quotes are uncommon, but they do occur — particularly when quantities are fractional or when a discount applies to only part of the quote. Always verify totals independently before sending. This is a two-minute check, not a reason to avoid the workflow.

What's the ROI on $20/month for ChatGPT Plus used specifically for quoting? If manual quoting currently costs you 30 minutes per quote and you produce 10 quotes a month, that's 5 hours of time. At a conservative $50/hour opportunity cost, that's $250/month in time spent on admin. Compressing that to 5 minutes per quote saves roughly 4 hours and $200/month — against a $20 tool cost. The math is straightforward. The Proposify data{:target="_blank"} adds a second dimension: proposals sent within 24 hours close at a higher rate than those sent later, so faster quoting has a direct revenue effect beyond the time saving.

Should I include my full price list in every prompt for security reasons? No. Avoid uploading documents that contain sensitive client data — tax IDs, banking details, or personally identifiable information. For the price list itself, the risk is low — you're not exposing confidential client information, just your own rate structure. That said, if your rate card contains commercially sensitive pricing you wouldn't want stored on a third-party server, use placeholder values in the document you upload and substitute real figures in your final document after generation.

Does this work for contractors and tradespeople, or just service businesses? It works for both, but contractor estimates need a few extra components that you should explicitly request in your prompt: a scope of work description, a materials-versus-labour cost breakdown, a validity period ("this estimate is valid for 30 days"), and a clear call to action. Build these into your master prompt template and the output will meet standard contractor estimate format without extra effort.