Sepsis results in more than 250,000 deaths each year in the United States. Resuscitation of the septic patient centers on timely recognition, early antibiotic administration, appropriate fluid administration, source control, and vasopressor administration for those with inadequate mean arterial pressures. At present, norepinephrine is the recommended first-line vasopressor for sepsis.  Current guidelines then recommend vasopressin as the second-line vasopressor agent, but the optimal timing of its initiation remains unknown. In this podcast, we discuss a recent article on the use of a machine learning tool for vasopressin administration to optimize short- and long-term outcomes in patients with sepsis.