
Quotes built on past quotes and on today's costs
Many customer requests resemble one another, yet each starts again from an empty spreadsheet, waiting for someone to say whether it can be done and at what cost.
Who prepares a quote today, and where it waits
The request arrives by email with a specification, or through the customer's procurement portal, and the sales office usually copies it into a spreadsheet before asking the engineering department or planning whether the materials and delivery dates hold up. In logistics the same goes for a transport request, to be compared with the carriers' rates. The time is lost in the waits between one department and the next and in the search for a similar past quote, which often exists in a colleague's mailbox.
Sales starts from a draft with today's costs and sets price and commitments
The agent reads the request with its attachments, extracts the specifications and deadlines and finds the closest matches among past quotes, with the price applied and the outcome. Using the feasibility criteria written in the Company Brain, which holds the company's rules, it shows what needs checking and by whom, then puts the draft together with up-to-date costs from the ERP system. The checks are signed off by the engineer or the planner; price, discounts and promised dates are decided by sales, because they commit the company to the customer.
- The request arrives, by email or through the customer's portal
- The agent extracts the specifications and finds similar quotes
- Prepares the draft and shows who has to check what
- Sales sets price and delivery and sends the quote
Illustrative example
Data, modules and decisions behind a quote
| Item | In this case |
|---|---|
| Data | In this casePast quotes with their outcome, price lists and costs from the ERP system, bills of materials and data sheets; in logistics, carriers' rates |
| Modules | In this caseData Foundation, Company Brain, Agents, Workspace |
| Who decides | In this caseSales on price, discounts and promised dates; the engineering department or planning on feasibility |
| What we measure | In this caseThe time between request and sending the quote, the waits between sales and the engineering department and the quotes redone because of an error, compared with the quotes produced before the agent |
| AI Act risk class | In this caseA quote to a business customer is a long way from the Annex III high-risk uses we meet most often, such as the creditworthiness of individuals; we check the classification all the same with those who look after risk and compliance, before writing the agent |
| Where to start | In this caseWith Deployment, when past quotes are collected somewhere and the head of sales wants to change the way of working |
Without past quotes to consult, the draft is worth little
When the quotes of recent years are scattered across the sales team's mailboxes, with no outcome or final price, the agent has no similar cases to start from. The same happens if the costs in the ERP system have not been updated for months, or if almost every request is a one-off, where the engineer's judgement is what counts. In these situations we begin by collecting quotes and costs, and build the agent afterwards.
The other cases for manufacturing are on the industry page.
Let's start again from a quote that arrived late
Taking a quote that arrived when the customer had already chosen another supplier, we work out with you which checks held it up.