Hospitality wholesaler (28 employees): price lists in 18 hours instead of 120

Example case · anonymized · figures are illustrative

The situation

A hospitality wholesaler in the Randstad, the urban west of the Netherlands, has 28 employees. More than 12,000 items from 40 suppliers: from tableware and glassware to cleaning supplies. Its clients are restaurants, caterers, and hotels that order through the online store and by email.

The company had one buyer who managed the prices. He knew every item and every supplier. That was his strength, and at the same time the risk: all that knowledge sat in his head and in one Excel file everyone called “the master list.”

The process before Optimi

Every quarter, and sometimes in between, it went like this:

1. A supplier emails a new price list. One sends a tidy Excel, another a 60-page PDF, a third a scan.
2. The buyer opens the list, looks up the supplier number for each item in the master list, and retypes the new purchase price.
3. He calculates the sales price with a formula in the master list and eyeballs it for anything odd.
4. The master list is exported and read into the ERP and the online store by hand.
5. In between come the backorders: every day, sales support writes about 25 emails to clients saying “your item X arrives on date Y.”

A full round of all suppliers cost around 120 hours a quarter. Three business days per list, sometimes more.

Where it failed

Two things. First, the errors: retyping thousands of lines, something went wrong now and then, and the client found it. Nine times a quarter, a complaint came in about a wrong price on the invoice. That cost money and trust.

Second, the dependency. If the buyer was sick, price processing stopped. Nobody else dared to work in the master list, because one wrong formula and the online store showed 12,000 wrong prices.

The backorder emails were not a disaster, but they were the same thing 25 times a day: look up the status in the WMS, write the email, send.

What Optimi did

Scan. Three processes surfaced in the scan: price-list processing, backorder communication, and the management of the master list itself. All three hung on the same Excel file.

Prototype. We started with the price lists of the five largest suppliers and built a list standardizer in ten business days.

What is in place now:

• An intake folder where price lists come in (by email or upload).
• An AI step that recognizes, for each list, which column is the item number, the description, and the price, whatever the layout, and links them to the company’s own item structure.
• A check that flags every price change above 15%, plus items that are new or missing.
• A review screen where the buyer looks at the flagged lines and approves the rest in one go.
• A bulk update to the ERP and the online store, with a log.
• For backorders: an agent that drafts an email per client from the WMS status. Sales support approves the whole list in one go each morning.

Custom apps and legacy bridge. The master list itself became a small web app with four users, roles, and a change log. The ERP stayed where it was; the app is the layer in between. The wholesaler owns the source code.

A person stays in the loop for: every flagged price change, every new item, and every backorder email (approved in bulk, but always approved).

The result

Indicative results for this scenario, measured over two quarters:

• Price-list processing: from 120 to 18 hours a quarter.
• Pricing errors reported by clients: from 9 to 1 a quarter.
• Backorder emails: from 25 a day written by hand to 0. They are all still sent, but as an approved draft.
• One critical Excel file replaced by an app that four people can operate.

The buyer still does the work that matters: negotiating, making the call on exceptions, assessing new suppliers. The retyping is gone.