This document approaches the critical need for accurate recession prediction in Bolivia by applying machine learning methodologies, specifically Logistic Regression, Random Forests, and Extreme Gradient Boosting. During the past three decades, disruptive events such as pandemics, financial crises, and geopolitical conflicts have highlighted the importance of early warning signals for anticipating economic downturns. However, forecasting recessions is complex due to the rarity of these events and the limited data available. Whereas traditional methods dominate existing literature, Bolivia lacks an official recession predictor. Our approach aims to identify turning points in economic activity through comprehensive data integration, providing a more accurate predictor than conventional methods. We found that real, monetary, and fiscal variables are relevant for predicting this indicator. Even though the findings are not definitive, they contribute to the empirical literature and provide a foundation for future research in this field, eventually assisting policymakers in mitigating the impact of economic recessions.