Developing drug substances with improved efficacy and reduced side effects remains a central challenge in lead generation and optimization. This poster explores how AI driven computational solutions are being applied to optimize potency, selectivity, and developability simultaneously. Key challenges include balancing multi parameter optimization, managing trade-offs between efficacy and safety, and translating in silico predictions into experimentally validated improvements. The session will highlight how AI models can guide chemistry and biology decisions, reduce iteration cycles, and prioritize high quality leads. Discussion will focus on practical implementation, data limitations, and integration with experimental workflows to accelerate optimization outcomes.