Poradnik
AI in quoting: what to automate so salespeople have more time for clients
With AI in quoting it is worth starting from tidying up data and repetitive steps: collecting parameters, preparing the description, listing items and producing the document for the client. Prices, calculations and commercial terms should follow set rules and current sources. That makes it possible to judge whether the tool really helps the sales team.
Preparing a quote often involves far more than writing text. A salesperson looks up information, compares variants, copies items across, keeps track of price-list versions and checks the document. This is a good starting point for a sales manager who wants to recover the team’s time without lowering the quality of service.
Which parts of quoting lend themselves to improvement
First split the process into three areas: gathering data, preparing the quote and approving it. In the first you need complete information from the client. In the second you combine it with the catalogue, prices and company rules. In the third you check whether the document can be sent.
AI can help organise an enquiry written in free language, or prepare a description based on approved information. Calculations that follow known rules are better handled by functions whose output can be checked unambiguously. Rights to discounts, exceptions and sending the quote are set separately.
You do not have to automate every variant at once. The first version can be limited to the most common category of quote, leaving unusual enquiries to the existing process. The tool must recognise the boundary of its scope and point out missing information.
A quote generator or an AI agent for sales?
A generator that produces the document once parameters are entered may be enough when the course of the task is predictable. An AI agent may be needed when it has to choose the next action, for example working out what data is missing and which available source to get it from.
Set the expected behaviour first. Does the system only create a draft? Does it save the quote in the CRM? May it send the quote itself? Each of those changes the scope of work and the responsibility. In a pilot, define precisely the point at which the salesperson approves the quote.
How to calculate the potential time saving
You need the number of quotes and the time to prepare comparable documents. Comparing a simple price estimate with a complex technical quote will not show the tool’s effect. Include checking and corrections in the measurement, even when someone else does them.
Worked example, all figures fictional: a company prepares 20 similar quotes a month. Each takes 90 minutes of work. After the improvement we assume 15 minutes, including checking.
| Value in the worked example | Before | After |
|---|---|---|
| Time per quote | 90 min | 15 min |
| Quotes per month | 20 | 20 |
| Total time | 30 hrs | 5 hrs |
The difference is 25 hours a month. That is time that can go to other work, if the rest of the process allows it to be used. It does not automatically mean a set percentage rise in sales.
Worked example · fictional figures
Human working time [hours / month]
25 hours of difference under these assumptions
This is not a measurement of a deployment, nor an automatic rise in sales.
Change the numbers above to work out your own case. The model assumes comparable quotes and time including checking.
Assuming an hourly cost of PLN 120, the value of that time would be PLN 3,000. The rate is fictional too. To assess the project you still have to add the cost of building, tools and maintenance. Freed time and a real reduction in company spending should be reported separately.
What to measure besides speed
Check the number of quotes needing correction, the time to reply to the client and how well items match the documentation. If the solution adds a drawing, a comparison of variants or a clearer description, judge how useful those are to the recipient as well.
Be careful about attributing causes when assessing sales. A change in the number of orders may come from price, season or the type of enquiry. Show first the effect that can be tied to the improvement: less work on the document, shorter waiting, fewer corrections.
What should remain after the pilot
The sales team should know the scope of quotes covered, the source of prices, how data is updated and who is accountable for the result. Agree what happens when the price list or catalogue changes, so the tool does not perpetuate old information.
At dyrektor.ai we combine understanding what salespeople need with choosing the scope and running the work. If the idea needs IT, it also helps to have a prototype ready to hand over to the technical team.