How Casa Vela Cut Food Waste 38% with AI Inventory Forecasting

The Challenge

Casa Vela is a chef-owned upscale restaurant in Phoenix with 80 covers and a menu built around fresh, seasonal ingredients. Inventory was ordered based on memory and instinct, the chef estimated what the week would need based on last week's feel. The result was a perpetual oscillation between too much and too little.

Over-ordering led to $4,000 per month in spoilage. Under-ordering meant 86'ing popular menu items on busy nights, the hospitality equivalent of turning away revenue at the door. Reservation volumes, local events, and even Phoenix weather patterns all affected demand, but none of these signals were being factored into purchasing decisions.

The Solution

Majoto built a demand forecasting model that treats the restaurant's order volume as a function of multiple inputs: reservation data, historical daily sales patterns, local event calendars, and weather forecasts. Each morning, the system generates a prep list and ordering recommendation calibrated to that specific day's predicted demand.

The chef reviews a one-page daily brief instead of making inventory guesses from scratch. The model learns over time, incorporating feedback when actuals diverge from predictions, continuously tightening its accuracy through each service period.

The Results

• 38% reduction in food waste
• $2,800 per month in spoilage savings
• Menu item availability on service up from 73% to 91%
• Chef's planning time redirected from inventory guesswork to creative development