Historically, a price catalogue (a mercuriale in French) recorded prices observed on a public market, for meat or agricultural produce, and served as an objective reference for negotiating purchasing budgets. Today the term refers to an internal or contractual document that lists items whose price has been negotiated. A product listed in a price catalogue has a fixed price for a given period, or one subject to variation rules agreed in advance.
What a price catalogue is for
Ideally, a price catalogue is a central tool for the buyer, for three reasons:
- Its simple format makes it quick to check that negotiated terms are being applied, over the year or according to the quantities purchased.
- It gives a basis for controlling purchasing budgets, since the price of every item is known in advance.
- It serves as a centralised reference for enforcing procurement policy across every subsidiary and every user.
The problem is the word “ideally”. A catalogue of several thousand lines, revised every year, often lives in a PDF or a spreadsheet, and nobody has time to check it against every line of every invoice. Control is done by sampling, and discrepancies slip through. Checking rate schedules and price catalogues is precisely what AI makes practicable at scale.
What AI changes about using a price catalogue
Bringing artificial intelligence into the handling of a price catalogue turns a static document into an active reference, continuously confronted with purchasing documents. Provided you choose a solution designed for the numerical data a procurement department handles: unit prices, units, discounts, volume conditions.
Detecting price discrepancies
The AI reads the catalogue, whatever its layout, then compares each invoice line received with the corresponding negotiated price. It flags pricing inconsistencies, unit mismatches and volume discounts not applied. Every discrepancy comes with its evidence: the catalogue line and the invoice line, side by side, with the amount at stake. The buyer then decides whether to dispute, request a credit note or let it go; the tool contacts no supplier on their behalf.
Negotiations prepared with data
By confronting quotes, purchase orders and invoices with catalogue prices, the AI identifies the items where terms no longer hold: increases applied without an amendment, suppliers invoicing outside the schedule, products bought at a higher price than an equivalent reference. That history prepares the negotiation with facts, before the contract expires rather than after. Price increases show up on the first affected invoice, not in the annual review.
Sourcing backed by history
By cross-referencing the catalogue’s prices and volume conditions with archived quotes and invoices, the AI surfaces the items for which another supplier, already consulted, offered better. It is decision support, not an automatic recommendation: competitive tendering remains the buyer’s job.
A shorter purchasing cycle
Decisions come faster because the procurement team works with up-to-date, already compared data. Whatever the format of supplier documents (quote, purchase order, invoice), the AI normalises them for consistent processing. The time saved goes into extending the perimeter: multi-site data, price history, quotes that were never processed.
What AI does not do
A serious solution does not promise margin that appears on its own. It changes nothing in the ERP, commits no payment, writes to no supplier and trains no model on your data. It investigates each discrepancy, documents it, and leaves the buyer to decide. On the perimeters we have analysed, recovered margin sits between 1 and 7% of the analysed perimeter: an observed bound, not a promise.
Time for strategy
By automating transactional processing and repetitive analysis, AI gives time back to buyers, from 15 to 45% per FTE, depending on perimeter and data maturity. That time goes to strategic thinking, supplier relationships and negotiation, where the buyer creates value. The price catalogue stops being a document consulted in case of doubt and becomes the rule every invoice has to respect.
To measure what your price catalogues let through today, the 3-week diagnostic on your data confronts your invoices with your negotiated prices and shows you the discrepancies, line by line.
