Build weekly vendor order sheets from par levels

Turn your inventory count into a suggested order for each vendor so ordering takes minutes, not an hour.

The problem

Ordering is usually done from memory or a clipboard after close. The result is running out of key items mid-weekend, over-ordering produce that spoils, and an hour or more each week spent on calls and order guides.

Step by step

1. 1Make a sheet with every item you order: item, vendor, unit, par level (how much you want on hand after delivery) and current count. If you use MarketMan or a similar inventory app, export your counts from it.
2. 2Count on the same day each week and fill in the current count column (a phone and Google Sheets works fine on the line).
3. 3Paste the sheet into ChatGPT, or use Copilot in Excel, and ask for an order per vendor: par minus on-hand, rounded up to the vendor's case size.
4. 4Adjust for anything unusual (a holiday, a private party, a slow week coming) and send each vendor's list through their portal or by email.
5. 5After a month, ask AI to compare what you ordered with what you threw out and suggest par changes.

Copy-paste prompt

Here is my restaurant inventory sheet with columns: item, vendor, unit, case size, par level, current count. [paste sheet]. For each vendor, build an order list: quantity = par minus current count, rounded up to the case size, skipping anything at or above par. Show a table per vendor with item, quantity in cases, and units. Then list any items where current count is below 25% of par so I can check them first. Do not change par levels; just flag anything that looks wrong.

Review AI output before using it. Don’t paste confidential, customer or patient data into tools your company hasn’t approved.