Industrial companies: answering RFQs faster with AI

In many industrial companies, a quote goes through one or two people who know the prices, the machine rates and the suppliers. When they are on leave, on site or simply overloaded, RFQs pile up. Meanwhile the client waits, and has often sent the same RFQ to several of your competitors.

At Oliverlist, AI sorts prospect replies as they come in, because a request that waits goes cold. An RFQ follows the same rule, with far more engineering involved: someone has to read drawings, find prices and check the workshop's capacity. Manufacturing is not my trade, and your estimator remains essential. But part of their time goes on searching and formatting, and that is the part AI can take on.

Here is what it can do, step by step. These are possibilities to adapt to your business, not client references.

Reading the RFQ or the drawing

An RFQ rarely arrives complete, and never in the same format: an email with PDF drawings, a specification, a BOQ in Excel, a photo of a worn part, notes from a phone call, sometimes in two languages.

AI can read the request and turn it into a quoting sheet:

• the items or services requested, quantities, delivery schedule;
• material, dimensions, finish, tolerances and the standards cited;
• requested delivery date and site;
• what is missing before you can price it.

That last point matters as much as the rest. A missing quantity, an unclear material grade, an unspecified coating: AI can draft the clarification question to the client straight away, instead of letting the RFQ sit until someone reads it.

Pulling prices and lead times

A quote rarely starts from zero. You have already priced similar parts, and you have supplier price lists, machine rates and routings. The problem is that this information is scattered: in the ERP, in spreadsheets, in email threads and in your estimator's head. Supplier prices also come in several currencies, with freight and duties on top.

AI can:

• find past quotes for similar items, with the price offered and the outcome, won or lost;
• pull the latest price from each supplier, with its date and currency, and flag a price that is too old;
• convert currencies at the rate you set, not at a rate it guesses;
• suggest manufacturing times from existing routings;
• read the production plan to propose a realistic delivery date, provided the plan is up to date.

All of this depends on accessible data. If past quotes are filed somewhere and supplier prices arrive by email, the project is realistic. If everything lives in one person's head, the first job is to write it down. AI can help by interviewing your estimator and structuring the answers. It will not guess the prices.

A draft quote, approved by a person

With the sheet and the prices, AI assembles a draft: line items, quantities, unit prices, lead times, terms. Every price shows its source: this past quote, this supplier price list from this date, this routing time. When it finds no source, it leaves the line blank and says so, rather than inventing a plausible number.

Your estimator reviews, corrects, sets the margin and approves. The commercial decision stays with them: a discount, priority for a key account, the choice not to bid. AI spares them the searching and the formatting. It does not set the price.

To know whether it works, measure two things before and after: the time between receiving the RFQ and sending the quote, and the number of quotes sent per week. Your win rate will come later, and it depends on much more than speed. I explain the method in Measuring the ROI of AI without inventing numbers.

Following up supplier orders

Once the quote is won, the work continues: you buy materials and components, then make sure suppliers deliver. With suppliers in Europe and Asia, lead times are long and shipments need watching.

AI can read order acknowledgements, compare them with your purchase orders (price, quantity, dates), flag differences and draft chasers to late suppliers. Above all, it can warn your planner when a supplier delay puts the client's delivery date at risk, before the client finds out. On project work, where late delivery can trigger penalties, that warning matters.

This is where quoting and delivery meet. At Oliverlist, automation has covered our whole chain since 2020, from importing contacts to issuing invoices. What I took from it: the gain comes from the whole chain, not from one isolated step. In industry, the RFQ, the quote, the supplier order and the delivery form a single chain. Quoting faster without tracking suppliers only means promising dates sooner, and keeping them less often.

Key takeaways

• Start where requests wait: reading the RFQ and finding past prices.
• AI prepares, your estimator decides on price and margin. No quote goes out without approval.
• Every price shows its source. No source, no number.
• The prerequisite is accessible past quotes and price lists. If they only exist in someone's head, write them down first.
• Measure the time from RFQ to quote, before and after.