CaterWiz: turning catering math into a system.

Where it stands

In pilot with Paheli Modern Indian Cuisine in Wayne, New Jersey. We will publish results once the pilot has produced them.

The problem

Most kitchens work out a catering order by hand. Someone scales each recipe to the guest count, estimates how much to buy, and finds out the margin after the event is over.

One missed conversion or one forgotten ingredient becomes an emergency trip to the store, or a job that did not make the money it should have.

The solution

CaterWiz starts from what the kitchen already knows: its menu, its recipes, and the order in front of it. From the guest count and the menu choices it works out exact prep quantities, builds one shopping list grouped by vendor, and shows food cost and margin before anything is bought.

Each order keeps a snapshot of the recipes and prices it was built on. Changing a recipe later does not rewrite the numbers of an event that has already happened.

Prices come from the kitchen's own dated purchase history. CaterWiz does not present them as live quotes, and it labels an estimate as an estimate.

For restaurants on Clover, the TillIQ integration adds read-only insight into sales, orders and payments next to the catering figures. Owners can also connect CaterWiz to an AI assistant such as ChatGPT and ask about their orders, shopping lists and costs in plain English. Both are optional: every catering feature works without them.

What we built

• Catering orders and quotes built from guest count and menu choices
• Versioned recipes with yields, waste factors and serving formats
• Shopping lists grouped by vendor, rounded to the package sizes actually sold
• Vendor price history drawn from the kitchen's own purchases
• Food-cost and margin analysis for each order and across orders
• A review queue for imported menus and receipts, so nothing changes until a person confirms it
• TillIQ: read-only Clover integration for sales, order and payment insight
• A connection to AI assistants for plain-English questions about the business

What this shows

This is the kind of problem we like best: a business running an important process on memory and spreadsheets, with real money depending on the arithmetic.

It also shows how we use AI. The assistant answers questions from the kitchen's own records. It does not invent prices, and it does not change anything a person has not approved.

← All case studies