Using AI to Compare Supplier Quotes: Make Them Comparable First

Make the quotes comparable first. Agree the fields, keep each value next to where it came from, and flag what’s missing. AI can help pull out the details, but a clean table is a starting point, not a supplier decision.

Can’t every supplier just use one template?

Try it, and plan for the ones who won’t. The problem is comparability, not just typing the numbers in.

1. Agree the fields before extracting anything.

Item and quantity. Match the specification, item revision, unit and order quantity.

Commercial terms. Capture currency, quantity breaks, minimum order, freight, payment terms and how long the quote is valid.

Delivery and exceptions. Keep lead-time assumptions, exclusions and substitutions visible.

Practical check: A blank field is not agreement to your preferred term.

A fictional example: same number, different basis.

Supplier A: CAD $10 per unit, freight included.
Supplier B: CAD $80 per box of 10, freight not stated.
Comparison: B is $8 per unit before freight, so there is no total-cost winner yet.

Practical check: Check the specification, quantity and delivery before ranking anything.

2. Keep the source beside the value.

Original. Keep the exact quoted value, and the document or passage it came from.

Normalized. Show the conversion, rate or unit rule used for the comparison.

Uncertain. Flag missing or conflicting terms, and prepare a question for the supplier. Don’t fill the gap with a guess.

Practical check: Confidence in the extraction is not commercial approval.

3. Test the difficult cases.

Include a revised quote, an expired offer, a different pack size, missing freight and an alternative specification. Check that rows line up and which version of the quote is being used. Verify the calculations yourself.

Practical check: Choosing the supplier and committing to buy stay with the authorized process.

The example is fictional. This is general guidance, not advice for your situation.