Prescience

Prescience operates an employer-sponsored health plan that acts as both the payer and the provider, letting it change what care members receive rather than just adjudicating claims. It is aimed at companies buying benefits for their teams, and the pitch is a plan built around an AI decision system rather than around the incentives of a traditional insurance carrier. The launch matters because the founders are showing the underlying model publicly for the first time, framing coverage design and clinical choices as a single optimization problem. The system has two parts described in the launch. Time Machine is a representation model that compresses a member's longitudinal health data into candidate future trajectories, and Crystal Orb is a dynamics and prediction layer that scores sequences of clinical actions against long-run health outcomes. The team illustrates the approach with a recurrent UTI case where the model preferred a narrower, older antibiotic over Cipro to preserve future efficacy, a move most urgent-care workflows would not make because they optimize for short-term symptom resolution rather than years-out risk like sepsis from resistant infection. Prescience is a Y Combinator company based in San Francisco, founded by brothers Rishab Jain and Aditya Jain, MD, who come out of Harvard and Harvard Medical School with backgrounds in machine learning and clinical medicine. For founders and operators evaluating benefits, the relevant question this launch raises is whether an integrated payer-provider running a learned policy can actually outperform the status quo on both member outcomes and employer spend.