Case Study: Compass India achieves 92%+ meal demand forecasting accuracy with Appinventiv's AI-powered planning platform

Compass India achieves 92%+ meal demand forecast accuracy with Appinventiv

The customer, Compass India, a food services enterprise operating across diverse sites, faced the challenge of accurately forecasting next-day meal demand. Their reliance on past averages and local judgment led to costly overestimates that created waste and underestimates that disrupted service and procurement. They engaged the vendor, Appinventiv, to build a custom AI-driven demand forecasting system to address this.

Appinventiv implemented a solution using multiple machine learning models, including Prophet and AutoGluon, to create site-specific forecasts. A key feature was a confidence scoring system that automated high-confidence predictions while flagging low-confidence ones for human review, with all overrides feeding back into the system for continuous learning. This approach resulted in over 92% forecast accuracy in pilot deployments for Compass India, significantly reducing waste and improving procurement planning.

Compass India