Combining machine learning and robotic precision, researchers present an integrated strategy for computer-augmented chemical synthesis, one that successfully yielded 15 different medicinally related small molecules, they say. Their novel, AI-informed, robotically controlled platform has the potential to greatly improve target-oriented continuously flowing chemical reactions and represents an important step towards fully automated and scalable synthesis of complex molecules. The design and synthesis of complex organic molecules are central to the discovery of useful novel compounds, including small-molecule pharmaceuticals. However, despite advances in laboratory automation, the synthesis of complex organic molecules largely remains a manual process, requiring considerable time and effort from chemists who oversee several labor-intensive steps. As such, an automated platform for chemical synthesis that is capable of both charting synthetic pathways and performing the flow chemistry required to produce large numbers of novel molecules is highly sought. Advancements towards such a system have largely progressed along two parallel tracks; some techniques have shown success in leveraging AI in compound design, while others make use of automated processes in reaction execution and production. Conner Coley and colleagues describe an open-source experimental system that integrates these two techniques. The platform's retrosynthesis algorithm is capable of generalizing millions of previously published reactions and of doing in silico validation, to propose successful synthetic routes. This data, once refined by human chemists, is compiled into reusable chemical "recipes," which are run on a modular flow chemistry platform that uses a robotic arm to automatically reconfigure and set up the unit to carry out the reaction. Coley et al. demonstrate their computer-augmented process by successfully synthesizing 15 different medicinally related small molecules with increasing complexity.