RFQ template for suppliers, written with AI
A good RFQ template for suppliers pins down five things: what you need, how many and by when, the commercial terms, the documents, and the exact format for the reply. AI can draft it fast and flag missing fields, but it must not invent data. Below are the template fields, a prompt and a quote-check table.
For many small businesses, a request for quotation is a two-line email: "Could you send us a price for this?" Each supplier then answers on its own terms, with a different quantity, delivery basis and validity period, and the three quotes on your desk can no longer be compared fairly. The RFQ template below prevents that. It asks every supplier the same questions, and AI helps you spot what is missing before the request goes out.
1. what a good RFQ contains
An RFQ works when a supplier can price the item without a round of follow-up questions, and when the answer lines up with the other quotes. That takes five blocks:
- what you need: description, specification, drawing or part number with its revision, acceptable alternatives;
- how many and when: first order quantity, expected repeat or annual volume, required delivery date, samples;
- commercial frame: currency, net prices, delivery basis (for example a named Incoterms rule and place), payment terms, quote validity;
- documents: required certificates, datasheets, declarations of conformity, warranty;
- how to respond: deadline, contact person, a fixed response format with unit price, MOQ and lead time on separate lines.
Leave any block out and the supplier fills the gap with its own assumption. That is how the cheapest-looking quote turns out to exclude freight, or to be based on a minimum order quantity (MOQ) ten times what you planned.
2. the RFQ template for suppliers
Fill it in first as an internal checklist. A field should stay blank only when you deliberately want the supplier to propose something.
supplier RFQ: fields
- subject: request for quotation – [product / service] – reply by [date]
- short introduction: who is asking and for what use (no confidential detail)
- item(s): name, specification, drawing number and revision, acceptable alternative: [fill in]
- quantity: first order, expected repeat volume, sample quantity: [fill in]
- dates: samples, first delivery, required lead time: [fill in]
- delivery: destination region, requested basis (e.g. Incoterms rule + place), packaging: [fill in]
- commercial terms: currency, net unit price, payment terms, validity: [fill in]
- documents: certificates, datasheets, warranty: [fill in]
- response format: unit price per quantity break, MOQ, lead time, freight on its own line, list of exceptions
- contact and deadline for questions: [fill in]
The "list of exceptions" line matters more than it looks. If a supplier cannot meet a requirement, it should say so there, not in the small print. When you compare, you immediately see who quoted in full and who quoted in part.
3. where AI helps, and where it should not guess
An AI assistant is not picking the supplier here. It is your editor and checker. It is useful for three jobs: turning rough notes into a structured RFQ, flagging missing or ambiguous fields, and producing the same request in several languages. A prompt that works:
- "Turn the notes below into a supplier RFQ in English and German. Use these blocks: item, quantity, dates, delivery, commercial terms, documents, response format. Do not invent any data: anything missing from my notes goes on a separate 'to clarify' list as a question. Also flag any point a supplier could read two ways."
"Do not invent any data" is not politeness. Language models tend to fill gaps with plausible details nobody asked for, such as a payment term that sounds standard. Before sending, check that every number and condition actually came from you.
Only upload confidential drawings, prices or customer data to a tool your company has approved for that. An anonymised description is usually enough to draft the RFQ.
4. checking the quotes that come back
A good RFQ is half the job. The other half is putting the answers into the same structure. AI can extract the fields from each quote and show where data is missing, but always check the figures against the original documents.
The table below is an illustration only. A, B and C are not real companies and the values are not market data.
| field | quote A | quote B | quote C |
|---|---|---|---|
| currency, net/gross | EUR, net | HUF, not stated | EUR, net |
| delivery basis | named rule + place | missing | ex works |
| MOQ | below requested qty | above requested qty | missing |
| lead time | stated | stated | "to be agreed" |
| validity | stated | missing | stated |
| exceptions | none | one item substituted | not marked |
"Missing" does not mean "meets the requirement". Send a short clarifying question for those fields and compare prices only once the answers are complete. If delivery terms differ, add freight and duties separately before ranking anything.
5. common RFQ mistakes
- no response format: every supplier structures the price differently, and comparison becomes manual work;
- no quantity breaks: you get one price for one volume, although repeat orders may price differently;
- vague delivery: "including shipping" is not a term; name the rule and the place (the official description of the Incoterms rules is on the ICC website);
- oversharing: the first round rarely needs the full drawing or your customer's name;
- no deadline: replies trickle in for weeks, and the earliest quote expires before you decide.
An RFQ only makes sense once you know who should receive it. If you are still building the list, our guide to AI vendor discovery walks through the brief, the shortlist and the company checks step by step.
frequently asked questions
what is the difference between an RFQ, RFI and RFP?
In short: an RFI gathers information about the market, an RFP asks for a proposed solution to a more complex need, and an RFQ asks for a price on a clearly defined item. This template is for the last one.
how many suppliers should get the RFQ?
As many as you can properly review and compare. Sending every supplier the same request matters more than the headcount.
can AI send the RFQs for us?
It can be part of an approved workflow, but a person should check recipients, content and shared data first. Drafts and a checklist are the safer place to start.
which language should the RFQ be in?
The supplier's. AI translates well, but check technical terms and units, because a wrong translation can describe a different product.
an RFQ template you will actually reuse
If requesting quotes, collating replies and following up still takes a lot of manual work, we help turn it into a repeatable AI workflow with human approval at every outgoing step. On the sales side, our article on AI sales meeting preparation uses a similar method. To shape it around your own purchasing, talk to oiSystem, or see what we do on the homepage.