Patients with cancer often undergo costly therapy and acute care utilization that is discordant with their wishes, particularly at the end of life. Early serious illness conversations (SIC) improve goal-concordant care, and accurate prognostication is critical to inform the timing and content of these discussions. This study uses a stepped-wedge, cluster randomized trial to evaluate the effect of a health system initiative using machine learning algorithms and behavioral nudges to prompt oncologists to have serious illness conversations with patients at high-risk of short-term mortality. Oncology practices will be randomly assigned in sequential four-week blocks to receive the intervention.