Steering high-throughput de novo design toward synthesis-ready leads with an agentic co-scientist

The challenge in modern drug discovery isn’t generating millions of ideas, it’s filtering that massive scale to find molecules that meet exact project criteria. Traditional lead optimization routinely gets trapped between two extremes: incremental edits that miss novel IP opportunities, or generative ideas that prove synthetically impossible in the wet lab. Bridging this divide requires keeping the chemist in control of an integrated, high-throughput architecture.

In this webinar, we explore how Schrödinger’s platform brings AutoDesigner, FEP+, and RetroSynth together inside Bunsen, our agentic co-scientist. Join Schrödinger scientists as they demonstrate how to steer complex, multi-step de novo design workflows through a human-guided agentic loop, ensuring every AI-generated idea is optimized for precise properties and ready for synthesis.

You will hear how Bunsen’s native toolset enables scientists to operate across isolated, multi-tasking sessions to run, monitor, and analyze parallel research streams without losing project context. The value of this workflow is already proven internally: within months of deployment across Schrödinger’s active discovery pipelines, roughly 80% of all computational jobs are now voluntarily routed through Bunsen. Join us to learn how to break the compromise between novel design and synthetic feasibility using human-guided agentic workflows.