Agentic AI: What it can and can't do for physicians

Agentic AI represents a new frontier in artificial intelligence, designed not just to analyze data or generate predictions, but to act autonomously on behalf of users within a set of defined goals. In health care, this means AI systems could carry out complex, multi-step tasks—such as scheduling follow-ups, drafting prior authorization requests, or coordinating care—without requiring constant human oversight.

Unlike traditional AI tools, which assist clinicians by offering insights or automation in narrow domains, agentic AI models operate with a degree of decision-making power. They can prioritize actions, adapt to changing inputs, and initiate workflows—much like a highly trained assistant who understands both clinical priorities and administrative protocols.

While it holds a lot of promise, agentic AI also raises important questions around safety, oversight, accountability, and trust. Medical Economics spoke with Isaac Park, CEO of Keebler Health, to learn more about the technology and how it can be used in health care.