How do you automate quotes with AI?

Where does the time go when a quote is put together?

Look at what a salesperson's day actually looks like. A large part of it is not spent with the customer at all.

It goes into digging out last year's quote to remember what you actually offered. Stitching together a new quote in Word for two hours. Trying to remember which of thirty contacts was already waiting for an answer last week.

All the work around the relationship that has to be done, but that no customer ever thanked a salesperson for.

That is where AI support belongs. Not in front of the customer. Next to the salesperson, on everything that eats the time between conversations.

In most B2B companies, quote time goes into the same things:

• Interpreting the request and understanding what the customer really wants.
• Asking for what is missing, often in several rounds.
• Finding similar quotes, prices and references.
• Writing and formatting the quote so it looks right.
• Getting the right person to approve price and terms.

What does the AI support do, and what does the salesperson do?

Quote automation moves the salesperson's time from compiling to judgement.

How does quote automation work, step by step?

1. The request is received. A forwarded email, an attachment or a note from a phone call is enough to start from.
2. The need is structured. The AI support picks out what matters for your quote: products, quantities, requirements, timeline.
3. Gaps are spotted. If something important is missing, the AI support suggests questions to the customer before anyone starts pricing.
4. History is matched. Similar past deals, prices and reference cases are pulled from your own knowledge base.
5. A draft is written. In your quote template, with your tone and your terms.
6. The salesperson approves. Adjusts in dialogue with the AI support and sends when the quote is right.

Why should the AI support never set the price itself?

Because price is a business decision. It rests on margin, the customer relationship and where the negotiation stands, and it must always be explainable.

So the AI support fetches prices from your pricing system or price list. If there is no verified price, it flags the gap instead of guessing. The salesperson, or whoever owns pricing decisions, always approves.

That is also what makes salespeople trust the tool. And that trust decides whether it gets used.

What do you need to get started?

Less than most people think. A first set of input is usually:

• Your current quote template.
• Five to ten quotes you have sent, showing tone and structure.
• A price list or access to the pricing system.
• A few reference cases you like to highlight.
• A salesperson who owns the process and tests it on real requests.

Integration with your ERP, CRM or pricing tool can come at a later step. You can start with a manual bridge where the salesperson pastes in the price, and connect the systems once the impact is proven.

How do you measure the impact of quote automation?

Measure before you start, and measure the same things afterwards:

• Lead time: time from request to quote sent.
• Effort: hours per quote.
• Volume: quotes per salesperson and week.
• Response and win rate: are the quotes getting better, not just faster?

That is how we work with the Prove & expand method: one process, one baseline, the impact in numbers before the next step. Read more in the guide How do you go from AI pilot to AI in production?

Example: Proposo

Proposo is our own solution for quote automation. The agent analyses needs, products, prices and reference cases and produces a structured quote draft in minutes instead of days. The salesperson adjusts and approves before anything is sent.

Proposo is an example, not a requirement. The same way of working can be set up in the quoting tool you already have. See more solutions in the portfolio.

What about customer data and regulation?

Quotes often contain personal data and trade secrets. So the data flows are mapped from the start: where data is stored, which models are used and who has access. Internal quote support where the salesperson always reviews is normally not one of the high-risk areas in the EU AI Act, but it should be documented. See also What is AI transformation? and our service ladder.

Sources: Regulation (EU) 2024/1689 (AI Act), EUR-Lex