Insilico, Lilly Unveil Vision for Prompt-to-Drug R&D

Got story updates? Submit your updates here. ›

Researchers from Insilico Medicine and Lilly have published a landmark perspective in ACS Central Science describing a comprehensive framework for fully autonomous, AI-orchestrated drug discovery. The paper outlines how an advanced reasoning system can integrate AI-driven target discovery, generative chemistry, automated synthesis, biological validation, and clinical planning into a single workflow, allowing a scientist to simply request "Design a drug for idiopathic pulmonary fibrosis" and have a central AI controller autonomously coordinate the entire drug discovery process.

Why it matters

The convergence of generative AI, multimodal foundation models, and automated laboratory systems is accelerating a fundamental transformation in drug discovery, but most pharmaceutical R&D remains fragmented across computational tools and manual experimentation. This framework aims to address that challenge by integrating all the key steps of drug discovery into a single, AI-orchestrated workflow.

The details

The paper traces the evolution of AI in biotechnology, highlighting how each step - from traditional machine learning to deep learning and transformer-based generative models - expanded AI's capability across target identification, molecular design, clinical prediction, and automated experimentation. It presents a next-generation architecture composed of sequentially called subsystems that are each autonomously driven and orchestrated by AI algorithms, including biology modules to mine data and validate targets, chemistry modules to design and optimize compounds, and clinical development modules to forecast trial outcomes and design strategies. The paper also emphasizes the importance of humanoid-in-the-loop automation to interact with legacy laboratory systems and enable uninterrupted 24/7 experimentation.

• The paper was published on February 20, 2026.

What’s next

The paper presents a conceptual framework, and the authors note that smaller, individual steps along the pipeline have already been automated and offloaded to AI-based programs in proof-of-concept studies. Bringing this vision to full fruition will require collaboration across the industry to integrate all the necessary components.