Agentic AI for Automated Drug Discovery: A Closed-Loop System for Hypothesis Generation and Testing

ABSTRACT

The traditional drug discovery pipeline is hindered by siloed computational predictions and slow, manual wet-lab validation. We present a fully autonomous, closed-loop Agentic AI system that seamlessly integrates hypothesis generation with robotic experimentation. Our orchestrator agent continuously analyzes genomic literature, designs novel molecular compounds via generative models, and directly commands cloud-connected automated laboratories to synthesize and assay the targets. In a 60-day continuous run, the system successfully identified and validated three novel kinase inhibitors, demonstrating a paradigm shift in pharmaceutical research velocity.