The application of AI in drug discovery relies on high-quality, reproducible training datasets. Traditional screening campaigns focus on identifying potent hits, but ML-driven drug discovery requires comprehensive potency evaluation across entire compound libraries.