We report the discovery of a novel, potent, and selective amidosulfonamide nonazapirone 5-HT1A agonist for the treatment of anxiety and depression, which is now in Phase III clinical trials for generalized anxiety disorder (GAD). The discovery of 20m (PRX-00023), N-{3-[4-(4-cyclohexylmethanesulfonylaminobutyl)piperazin-1-yl]phenyl}acetamide, and its backup compounds, followed a new paradigm, driving the entire discovery process with in silico methods and seamlessly integrating computational chemistry with medicinal chemistry, which led to a very rapid discovery timeline. The program reached clinical trials within less than 2 years from initiation, spending less than 6 months in lead optimization with only 31 compounds synthesized. In this paper we detail the entire discovery process, which started with modeling the 3D structure of 5-HT1A using the PREDICT methodology, and then performing in silico screening on that structure leading to the discovery of a 1 nM lead compound (8). The lead compound was optimized following a strategy devised based on in silico 3D models and realized through an in silico-driven optimization process, rapidly overcoming selectivity issues (affinity to 5-HT1A vs alpha1-adrenergic receptor) and potential cardiovascular issues (hERG binding), leading to a clinical compound. Finally we report key in vivo preclinical and Phase I clinical data for 20m tolerability, pharmacokinetics, and pharmacodynamics and show that these favorable results are a direct outcome of the properties that were ascribed to the compound during the rational structure-based discovery process. We believe that this is one of the first examples for a Phase III drug candidate that was discovered and optimized, from start to finish, using in silico model-based methods as the primary tool.