Economic growth, increasing global energy demand, and the depletion of known reserves are driving the oil and gas industry to explore deeper reservoirs. These deeper formations, particularly in brownfields, present unique challenges, with recovery factors often 10–30% lower than expected. Exploration Well A in the Z reservoir confirmed the presence of hydrocarbons but revealed significant challenges due to the reservoir’s low porosity and permeability. Limited Special Core Analysis (SCAL) data, caused by high costs, further complicates reservoir characterization and decision-making. Machine learning (ML), a branch of artificial intelligence, offers a groundbreaking approach to address these challenges. ML algorithms learn from data and adapt without explicit programming, making them highly effective in improving reservoir evaluation. By analyzing raw log data, ML models reduce reliance on fixed constants in traditional equations, leading to more accurate and consistent predictions. This study deployed ML techniques to predict porosity, permeability, and moveable fluid volumes in selected wells across the Sarawak Basin, Sabah Basin, and Malay Basin. The results demonstrated that ML predictions aligned well with conventional interpretations, with deviations consistently within acceptable thresholds. The technology also highlighted potential producible zones, including thinly laminated formations that conventional methods often overlook. The study emphasized the critical role of diverse, high-quality training datasets in addressing geological variations and borehole complexities. By integrating ML into reservoir workflows, the industry can enhance reservoir evaluation, reduce uncertainties in volumetric estimates, and support more reliable decision-making. This approach not only optimizes recovery strategies but also provides a path forward in managing the complexities of deeper reservoirs more effectively.