Artificial intelligence in restaurants: A solution to forecast demand and reduce food waste in Miraflores

This investigation assessed the impact of an artificial intelligence (AI)-powered demand prediction system on minimizing food waste in medium-scale restaurants located in Miraflores, Lima, Peru. Employing a quantitative pre-experimental approach with baseline and follow-up assessments, the study evaluated operational metrics from five establishments over a six-week period. The AI system, developed using machine learning techniques such as Random Forest and Gradient Boosting, generated daily projections of ingredient requirements by leveraging historical transaction records and situational factors. The findings revealed a significant reduction in daily food waste, decreasing from 4.12 kg to 2.76 kg after implementation. Statistical analyses indicated strong internal consistency (α = 0.841), normally distributed data, and a statistically significant difference in means (p < 0.001). These results highlight the value of AI-enhanced decision-making in optimizing resource use, reducing operational costs, and promoting sustainable practices within the food service industry. The study provides valuable empirical insights into digital innovation in hospitality, particularly relevant to the Latin American region.