---
title: Quote template with AI: from one day to 15 minutes | Efidis
description: The sheets, columns and fields of a quote template, the formula that pulls prices from a list, a prompt to have AI draft the quote and when it isn't worth it.
url: https://efidis.co/en/blog/plantilla-de-cotizacion-con-ia
lang: en
---

# Quote template with AI: from one day to 15 minutes

*Customers and sales · Efidis team · Updated November 12, 2026*

**In short:** A quote that takes a day shows up late: in a study of 1.25 million sales leads, companies that tried to respond within the first hour were nearly 7 times more likely to qualify the lead than those that waited even one more hour (Oldroyd, McElheran and Elkington, 2011). Getting down to about 15 minutes takes two pieces: a template with the price list on a separate sheet, so nobody types a price from memory, and an AI that drafts the quote from the customer's message. The rule that ties them together: AI writes, the list sets the price.

An illustrative example. A customer writes on Tuesday at 10 a.m.: "Can you quote me 40 cases of the large size and 12 of the medium, delivered Friday?" The quote goes out Wednesday afternoon. Nobody spent a full day writing it. It took 40 minutes of actual work, spread across looking up prices, asking the warehouse whether there was stock, getting a discount approved and waiting for someone to open the right file.

## Why speed matters more than formatting

The most cited study on response times was published in Harvard Business Review. The authors sent a test inquiry through the websites of 2,241 U.S. companies: 37% responded within an hour, 24% took more than 24 hours and 23% never responded at all. Among those that did respond within 30 days, the average was 42 hours (Oldroyd, McElheran and Elkington, 2011).

In a second part, they analyzed 1.25 million leads from 42 companies. Firms that tried to reach the customer within the first hour were nearly 7 times as likely to qualify the lead as those that waited one more hour, and more than 60 times as likely as those that waited 24 hours or longer (Oldroyd, McElheran and Elkington, 2011).

Read it carefully: the data predates 2011, it covers web inquiries, and "qualify" means getting a conversation, not closing a sale. Even so, the direction is clear: a customer's interest cools by the hour, not by the week.

There's a second reason. In 6sense's study of more than 4,000 business buyers, 94% of buying groups had already ranked their preferred vendors before first contact, and they ended up buying from that favorite 77% of the time (6sense, 2025). If you're not the favorite, your quote is your chance to change the ranking. If it shows up late, it just confirms what the customer already thought.

## The rule: AI writes, the list sets the price

Quoting has two slow parts. The first is finding data: prices, stock, lead times, the discount cap. The second is writing: figuring out what the customer asked for and sending back a clear answer. The template handles the first and AI handles the second. What can't happen is the two getting mixed up.

A language model writes well and confidently, even when it's wrong. In 2024, a Canadian tribunal ordered Air Canada to pay a customer because the chatbot on its website promised a discount the airline didn't offer: the company is responsible for what its chatbot says (Moffatt v. Air Canada, 2024). A quote works the same way. If the AI writes a price, that price is yours.

> The rule in one line: AI writes, the list sets the price. No number on the quote comes from the model's memory or a salesperson's. It comes from the price sheet, and the sheet does the math.

People who use AI every day already work this way: 86% treat what it produces as a starting point, not a final answer (Microsoft, 2026). The template turns that habit into a method.

## The template: three sheets, columns and fields

Build one file with three sheets: Prices, Quote and Log. One person keeps the first one up to date; salespeople fill in the second; the third tells you what's happening with everything you've quoted.

### Sheet 1: Prices

| Column | What goes in | Illustrative example |
|---|---|---|
| SKU | Unique, no spaces or special characters | BOX-040 |
| Product or service | The name as the customer should read it | Corrugated box, 16 × 12 in |
| Unit | Each, case, hour, foot, visit | Case |
| List price | Before tax | 12.50 |
| Max discount | What a salesperson can give without approval | 8% |
| Tax | The rate that applies where you sell | Depends on the item |
| Lead time | Usual business days | 3 |
| Updated | Date of the last price change | 2026-10-01 |

### Sheet 2: Quote, the header

| Field | What it's for |
|---|---|
| Quote number | So you can find it later and know how many you send a month. Never repeat one: take it from the Log sheet |
| Date and expiration | An expiration date (say, 15 days) protects you if prices change |
| Customer and contact | Company, person, phone or email |
| Salesperson | Who owns the quote |
| Delivery date | The one the customer asked for, or the one you can meet, written as a date, not "soon" |
| Payment terms | Due on receipt, Net 30, 50% deposit |
| Next step | What the customer has to do to accept: reply, sign, pay the deposit |

If you only need the document to send, without the spreadsheet, use the [PDF and Word quote template](https://efidis.co/en/blog/formato-de-cotizacion).

### Sheet 2: Quote, the line items

| Column | Who fills it in | How |
|---|---|---|
| SKU | Salesperson | Dropdown with the SKUs from Prices |
| Product | The sheet | Formula that looks up the SKU in Prices |
| Quantity | Salesperson | A number greater than zero |
| Unit price | The sheet | Formula; the cell is locked |
| Discount % | Salesperson | Validated: can't exceed the max discount |
| Subtotal | The sheet | Quantity × price × (1 − discount) |
| Tax | The sheet | Subtotal × the item's tax rate |
| Total | The sheet | Subtotal + tax |

The key formula is the one that pulls the price. In Excel 365, 2021 and 2024, and in Google Sheets: =XLOOKUP(A12,Prices!A:A,Prices!D:D,"Not on the list"). It looks up the SKU in cell A12 on the Prices sheet and returns its price; if the SKU doesn't exist, it says so instead of leaving a blank. Excel 2016 and 2019 don't have XLOOKUP, so use VLOOKUP there.

For the cap, add data validation to the Discount column so it won't accept anything above that SKU's max discount. If the customer asks for more, the salesperson knows they need approval; if not, they don't ask.

### Sheet 3: Log

One row per quote: number, date, customer, salesperson, total, time the request came in, time the quote went out, status (sent, won, lost) and reason lost. The two timestamps give you your real response time. After 30 quotes, the reason-lost column tells you whether you're losing on price, on delivery or on being late. The same sheet gives you the close rate the [Monday dashboard](https://efidis.co/en/blog/indicadores-semanales) asks for.

## Where the day goes, and where you get it back

Here's the breakdown for the opening example. The numbers are illustrative: track your own for a week before you change anything.

| Step | Before | With template and AI |
|---|---|---|
| Read the message and figure out what they want | 5 minutes | 1 minute: AI pulls out products and quantities |
| Look up prices | 15 minutes across files and emails | 0: the formula pulls them |
| Confirm stock | About 3 hours waiting on the warehouse | 3 minutes, if inventory lives in an up-to-date sheet or system |
| Approve the discount | Half a day waiting on the manager | 0 if it's within the cap |
| Write and send | 20 minutes | 3 minutes of AI and 8 of review |
| Total | 40 minutes of work and more than a day of waiting | About 15 minutes |

The work was never a full day; the waiting was. The list removes the lookup, the cap removes the approval and AI removes the writing. Stock is the one piece the template can't fix: if inventory lives in one person's head, that person is still the bottleneck.

## The prompt: how to have AI draft the quote

1. **Use a business account.** Open the AI tool on the business plan your company approved. OpenAI, for example, says that on its business plans it doesn't train its models on company data by default (OpenAI, 2026). Check that condition with your provider and put the rules in an [AI use policy](https://efidis.co/en/blog/politica-de-uso-de-ia).
2. **Give it the product list.** Attach the exported Prices sheet or paste it in full. SKUs, names, units and lead times only: it doesn't need prices to write.
3. **Paste the customer's message.** Exactly as it came in, typos and all. If it was a voice message, paste the transcript.
4. **Use the prompt below.** Save it as a template in the tool so you don't retype it every time.
5. **Move the lines to the sheet and review.** Copy SKU and quantity into the template, let the sheet do the math and check three things: that each product is what the customer asked for, the delivery date and whatever landed under "To confirm."

The prompt, to copy and adjust for your business:

- You are a sales assistant at [Company]. Below are our product list and a message from a customer.
- Pull out every product the customer asks for, with quantity and unit, and find its SKU on the list.
- If a product isn't on the list or the quantity isn't clear, don't quote it: put it under a "To confirm" section with the question we'd ask the customer.
- Don't write prices, discounts or taxes. Our spreadsheet calculates them.
- Return a table with SKU, product and quantity, and a two-paragraph email that confirms the delivery date the customer asked for and states the next step to accept.

Look at the third and fourth lines. They ask the AI to say "I don't know" instead of guessing, and they take the numbers away from it. That's the rule in practice: AI writes, the list sets the price. Then you copy the total into the email from the sheet, not from the AI's text.

## Checklist: eight things a quote that closes has

Before you send, check the quote against this list. If something is missing, the customer will have to ask, and every question costs another day.

- [ ] It has a quote number and a date, and it's in the Log.
- [ ] It says how long the prices are good for.
- [ ] Every line has a SKU, quantity and unit, and matches what the customer asked for.
- [ ] Prices come from the current list, and no discount goes over the cap without approval.
- [ ] Subtotal, tax and total are shown separately.
- [ ] It gives a specific delivery date you can actually meet.
- [ ] It states the payment terms: due on receipt, net terms or deposit.
- [ ] It ends with a clear next step: what the customer does to accept and who to reply to.

## When this method isn't worth it

- You send fewer than ten quotes a month. At that volume, a good template is enough and AI saves only a few minutes.
- Every quote is a project: a build, a custom job, consulting work that needs a site visit or discovery. The template still helps with fixed-price line items, but the value is in the assessment, and AI doesn't write that. If you quote construction jobs, what comes after acceptance is [construction cost control](https://efidis.co/en/blog/control-de-costos-de-obra).
- Your prices change daily and nobody updates the list. The AI and the formula will confidently read a stale list. Sort out who updates prices, and when, first.
- The customer requires their own format, as in a formal bid. Use the sheet to calculate and fill in their form by hand.
- You only have free AI accounts and the quote includes customer data. Use the template without AI until you have a business account.

> The template can be wrong too. In audits of real-world spreadsheets, 94% of the 88 spreadsheets reviewed across seven studies contained errors (Panko, 2005). Lock the formula cells, let only one person edit the Prices sheet and save a copy before every price change.

## When the spreadsheet stops being enough

The template starts to break when every salesperson keeps their own copy, stock gets checked over text messages and the Log is filled in from memory on Friday. At that point the problem isn't writing fast anymore: it's that the data behind the quote lives in different places.

The next step is having prices, inventory and customers in one system, so the quote is built from current data and gets recorded without anyone copying it. With the data in one place, connecting an AI to draft quotes is also safer: it reads the current list instead of a file exported last week. That's what we build as [custom software](https://efidis.co/en/servicios/software-a-la-medida), for example for [distributors](https://efidis.co/en/sectores/distribuidoras).

## Frequently asked questions

### What sheets should an Excel quote template have?

Three: Prices, with the current list and each SKU's max discount; Quote, where the salesperson enters SKU and quantity and the sheet pulls the price; and Log, with one row per quote sent. Prices come from the Prices sheet and are never typed in by hand.

### Can I use ChatGPT or another AI tool to write quotes?

Yes, for the writing: reading the customer's message, pulling out products and quantities and drafting the email. Not for prices. Tell it not to write numbers and to flag anything it can't find on your list as "to confirm," and let the spreadsheet do the math. Use a business account that doesn't train on your data.

### How fast should I send a quote?

As soon as you can, ideally the same day. In the study published in Harvard Business Review, companies that tried to respond within the first hour were nearly 7 times more likely to qualify the lead than those that waited one more hour (Oldroyd, McElheran and Elkington, 2011).

## Sources

1. [James B. Oldroyd, Kristina McElheran and David Elkington, "The Short Life of Online Sales Leads," Harvard Business Review, 2011](https://hbr.org/2011/03/the-short-life-of-online-sales-leads)
2. [6sense, "The Timeline for Influencing B2B Buyers Is Shrinking: Insights From 6sense's 2025 Buyer Experience Report," 2025](https://6sense.com/newsroom/the-timeline-for-influencing-b2b-buyers-is-shrinking-insights-from-6senses-2025-buyer-experience-report/)
3. [American Bar Association, "BC Tribunal Confirms Companies Remain Liable for Information Provided by AI Chatbot" (Moffatt v. Air Canada), 2024](https://www.americanbar.org/groups/business_law/resources/business-law-today/2024-february/bc-tribunal-confirms-companies-remain-liable-information-provided-ai-chatbot/)
4. [Microsoft, "2026 Work Trend Index," 2026](https://www.microsoft.com/en-us/worklab/work-trend-index/agents-human-agency-and-the-opportunity-for-every-organization)
5. [OpenAI, "Enterprise privacy at OpenAI," 2026](https://openai.com/enterprise-privacy/)
6. [Microsoft, "XLOOKUP function," Office support](https://support.microsoft.com/en-us/office/xlookup-function-b7fd680e-6d10-43e6-84f9-88eae8bf5929)
7. [Raymond R. Panko, "What We Know About Spreadsheet Errors," revised 2005 version](https://excel-do.com/wp-content/uploads/2023/09/What_We_Know_About_Spreadsheet_Errors.pdf)

## Send quotes the same day

Efidis builds your company's app so every quote goes out with current prices and stock. On the call we'll look at how your team quotes today and where the time goes.

- [Book my call](https://wa.me/573214646372?text=Hi%2C%20I%20read%20an%20article%20on%20the%20Efidis%20blog%20and%20want%20to%20see%20how%20to%20apply%20it%20in%20my%20company.)
