How to Automate Quotes and Estimates With AI - Send Them the Same Day

Why quotes are worth automating first

Quoting sits in an awkward spot. It is not the work, so it feels like it should be quick, but it needs your judgment, so it never quite is. The result for most small businesses is a pile of enquiries that sit for two or three days while you find an evening to write them up properly.

That delay is expensive in a way that never shows up on any report. The customer who emailed you on Tuesday is still interested on Tuesday. By Friday they have talked to two other people. You do not lose those jobs on price, you lose them on the calendar, and you never find out because nobody replies to say why they went elsewhere.

The Price Book is the whole thing

Here is the part people skip, and it is the part that decides whether this works or quietly costs you money. AI does not know what you charge. If you ask it to write a quote without telling it your prices, it will produce a number that looks completely reasonable and is entirely made up.

So before you automate anything, write a Price Book. It is one document, and it can be plain text. It holds the things you actually sell, what each one costs, and the rules you apply.

• Your standard jobs and the price or price range for each.
• Your hourly or day rate, and the minimum callout if you have one.
• The common add-ons and what each one adds.
• Your rules, written plainly. Things like travel beyond twenty miles adds a fee, or weekend work is charged at a higher rate.
• What is never included, so it does not quietly end up in a quote.

Writing this takes an afternoon and most people find it useful on its own, because it is often the first time their pricing has existed anywhere other than their head. Once it exists, you paste it in alongside every enquiry, and the quotes stop being guesses.

The Price Book, Draft, Check, Send loop

The whole system is four steps, and the third one is not optional.

1. Price Book. Your real numbers and rules in one document, kept up to date.
2. Draft. Paste the customer enquiry plus the Price Book and ask for a quote using only those prices, with anything unclear flagged as a question rather than assumed.
3. Check. You read the numbers, the scope, and the assumptions. Two minutes, every time.
4. Send. Out it goes, same day.

The instruction to flag rather than assume is what makes the check step fast. A draft that says I have assumed a single storey property, please confirm is genuinely useful. A draft that silently guessed is a trap.

What to check in those two minutes

That is a genuine two minutes once you have done it a few times, against the forty minutes writing the quote from scratch used to take. That is where the time comes back.

Handling the enquiries that are too vague to quote

A good share of enquiries do not contain enough information to price. The mistake is treating that as a failure of the automation. It is not, and it is a job the AI does well.

Ask it to produce a short reply with the three or four specific questions needed before you can quote, in a friendly tone. That reply goes out the same hour, which puts you in conversation with the customer while your competitors are still deciding whether the enquiry is worth chasing.

Rolling it out without breaking anything

1. Pick your single most common job type. Just one.
2. Write the Price Book section for that job only. Half a page is enough to start.
3. Run your next five enquiries of that type through it, and check every one properly.
4. Compare the drafts against what you would have quoted. Where they differ, the Price Book is usually missing a rule. Add it.
5. Once five in a row need no correction, add your second job type.

The Price Book gets better every time you correct it, and it is the asset here. The AI is interchangeable. Your written pricing rules are not, and they are useful for training anyone else in the business too, which pairs naturally with [how to train your team to use AI](/blog/how-to-train-your-team-to-use-ai).