Researchers used a near-autonomous AI chemist to improve a challenging reaction in medicinal chemistry, resulting in significantly higher yields for a key process. The AI system, GPT-5.4, worked with a high-throughput laboratory to design and run experiments, analyze data, and propose follow-up experiments. This collaboration led to a major breakthrough in Chan-Lam coupling, a reaction used to form carbon-nitrogen bonds, with human chemists validating the results. Why it matters: this achievement demonstrates the potential for AI to accelerate discoveries in medicinal chemistry, ultimately leading to more efficient drug development and benefiting humanity.