Reproducibility in chemical synthesis is challenging. This paper introduces DigiChemTree, an AI-backed platform that auto-optimizes photochemical reaction parameters for on-demand molecule synthesis. It uses a continuous flow reactor with integrated AI to optimize light-induced carbene generation and its coupling with various nucleophiles, achieving high yields and significantly reducing reaction times compared to traditional methods. The platform was successfully applied to late-stage functionalization of active pharmaceutical ingredients.

Abhilash Rana, Ruchi Chauhan, Amirreza Mottafegh, Dong Pyo Kim, Ajay K. Singh