This study investigates the advanced machine learning models' regression, XGBoost, and RNNs, for predicting the retailer's behavior concerning economic features to enhance pattern discovery. The two-year dataset from a major dairy company shows that the combination of economic indicators such as inflation rates and seasonal variations with machine learning models significantly improves prediction accuracy. The XGBoost model, RNNblended in particular, reveals higher-level intelligence in relating complicated retail patterns and, therefore, helps to manage inventory and optimize the supply chain better. Our results convey the power of adopting machine learning models with economic features, statistical models, purchasing patterns, and considering seasons and holidays in periods of the year actionable insights to organizations regarding customer demand prediction, inventory overstocking, and stockout reduction, thus, supporting marketing and operational efficiency improvement.