In short
- AI reads the enquiry, spots missing details, suggests items and prices from your price list and writes a draft.
- Prices always come from your own data, never from the language model. A person approves.
- The prerequisites are a maintained price list, text building blocks and clear rules for discounts and special cases.
- Success is measured by the time to quote and the share of drafts that go out without major changes.
Why quotes take so much time
In many small and mid-sized businesses the quote is a bottleneck. The enquiry arrives by email, phone or form, often incomplete. Someone has to read and understand it, ask questions, look up prices in lists or the ERP, assemble text from old quotes and transfer everything into the template. For individual services, the costing comes on top.
The time rarely goes into the actual professional decision, but into routine: typing up details, searching for item numbers, adjusting wording, following up later. That is exactly where AI comes in.
What AI can take over for quotes
- Read the enquiry: the AI extracts the key details from free text: customer, service, quantities, dimensions, dates, location.
- Spot gaps: if details needed for a quote are missing, it drafts a short follow-up question to the customer.
- Suggest items: from your catalogue or ERP it suggests suitable products and services with your prices.
- Write text: the service description and cover letter are created from your text blocks, matched to the enquiry.
- Create a draft: the quote is created as a draft in your system, not sent directly.
- Follow up: if there is no reply, the process sends a reminder after a set time and suggests a follow-up email.
Where people stay in charge
Prices, discounts, feasibility, delivery dates and special terms are decisions your business makes. The AI prepares and flags anything unusual, such as a very large quantity or a requested date that is getting tight. The responsible person checks the draft, changes it if needed and approves it. A quote is a binding statement by your business, which is why this approval is always part of it.
Quoting software, ChatGPT or your own process?
| Strength | Limit | Suits | |
|---|---|---|---|
| Quoting software | Clean templates, numbering, status | The groundwork from emails remains manual | Businesses with few, similar quotes |
| ChatGPT on its own | Fast text and wording | Does not know your prices, copy and paste remains | Occasional cover letters |
| Automated process with AI | Connects enquiry, price list and system | Needs maintained data and clear rules | Many enquiries, recurring services |
Often the best solution is a combination: your existing quoting or ERP software stays, and an automated process delivers the groundwork to it.
Prerequisites
- A maintained price list or catalogue in a system that can be connected, or at least a clean spreadsheet.
- Text blocks for your most common services, cover letters and terms.
- Clear rules for discounts, minimum quantities, surcharges and special cases.
- Access to customer data in the CRM or ERP.
- A template with all the details that belong in your quotes.
- A named person who approves, and a deputy.
A typical process from the trades
A plumbing business receives an enquiry through its form for a bathroom refit, with three photos and the sentence “shower instead of bath, about eight square metres”. The AI recognises the service, notices that details on tiles and the preferred date are missing and drafts a friendly follow-up question. After the reply, it suggests the items for removal, installation, tiling and fitting from the service catalogue, calculates with the stored prices and creates the draft. The master plumber checks it, adjusts two items and sends the quote. This example describes a typical process, not a client project.
Data protection and security
Enquiries contain names, addresses and sometimes photos from homes or premises. Processing them requires a data processing agreement, servers in the EU where possible and access rights that only go as far as the process needs. The AI reads prices and customer data but changes nothing without approval. Every step is logged so that it is clear later how a quote came about.
What you should measure
- Time to quote: from receipt of the enquiry to sending.
- Change rate: how many drafts are approved without major changes.
- Follow-up questions: how often customers still need to clarify something after the quote.
- Conversion rate: whether faster quotes lead to more orders.
More on processes like this under AI process optimisation and in our article on customer enquiries with AI.