This work presents an intelligent analytics framework that integrates advanced machine learning techniques, natural language processing, and multidimensional evaluation methodologies to optimize organizational talent management at leadership level. The proposed framework strategically employs different learning models to enhance situational awareness for organizational decision making process in an information technology institution, providing comprehensive insights into the performance metrics and improving the potential of subordinates occupying executive positions. The comprehensive methodology includes the implementation of robust classification algorithms (Random Forest, XGBoost), advanced clustering techniques (K-means), and sophisticated sentiment analysis using tools like a Bidirectional Encoder Representations from Transformers (BERT) to systematically process and analyze qualitative feedback from multiple organizational sources. The system undergoes rigorous validation through a carefully designed pilot study conducted in a technology organization, demonstrating exceptionally high psychometric reliability coefficients (alpha de Cronbach > 0.90) across all evaluated dimensions. Expected outcomes include automated identification and categorization of distinct talent profiles, accurate prediction of individual and organizational development needs, and significant optimization of strategic decision making processes. This innovative approach represents a substantial contribution to advancing evidence based talent management practices through sophisticated computational analysis and data driven organizational intelligence.