Artificial intelligence has arrived on both sides of the hospital-health plan negotiating table, and Deborah Greer, MD, wants to make sure one thing doesn’t get lost in the speed: the physician who has to own the call.
Dr. Greer is the system lead for advanced appeals and settlements, denials management and prevention at Chicago-based CommonSpirit Health, where her team sits at the intersection of clinical care and health plan coverage. Her job, and the job of the physician advisers she oversees, is to ensure patients are placed in the correct care status and to take up the conversation with health plans when there’s a disagreement. As AI tools proliferate on both the hospital and payer sides of that conversation, Dr. Greer is drawing a clear line about where the technology’s role ends.
“The guardrail I keep coming back to is that the decision has to stay with the physician who can be accountable for it,” she said in a recent interview with the “Becker’s Healthcare Podcast“. “If automation moves a determination faster without a clinician actually looking at the case, then all we’ve done is remove the judgment and keep the speed.”
The concern is not theoretical. Medicare Advantage plans made nearly 53 million prior authorization determinations in 2024, and AI tools are increasingly embedded in how those decisions get processed. Dr. Greer said the technology is genuinely useful; it can pull a complex chart together quickly, flag a documentation gap while the patient is still in the bed, and surface patterns across thousands of denials that no individual reviewer would ever see. That’s real value, but there’s also risk.
The governance challenge is compounded by the pace of adoption. Dr. Greer said the healthcare industry is correct to move carefully on tools that are still being validated, where performance benchmarks aren’t settled and the consequences of a wrong determination extend to both patients and organizations.
“You’re being asked to move quickly on something that’s largely untested in an environment where being wrong has consequences for patients and for healthcare organizations,” she said. “So the governance is still being written, and it has to catch up with what the technology can already do.”
That regulatory gap is drawing attention in Washington as well. In July 2026, bipartisan legislation was introduced in the House — the Protecting Patients from Automated Denials Act — that would require any AI-assisted Medicare Advantage denial to be independently reviewed and approved by a qualified physician, who would attest in writing that their judgment was independent of the algorithm’s output.
Dr. Greer sees a version of the technology that supports physician decision-making by organizing information and identifying gaps, but stops short of generating the determination itself.
“I believe AI can help prepare the argument, but it’s a physician that makes the clinical determination and owns it,” she said.
The stakes for getting that balance wrong are high on CommonSpirit’s scale. The 137-hospital system is in the midst of Project Impact, a sweeping operational overhaul targeting $5 billion to $6 billion in value creation over 36 months, with revenue cycle improvement identified as a central workstream. Denials management sits squarely in that effort and Dr. Greer sees AI’s promise in the field as real, provided the guardrails keep pace.
Beyond governance, she flagged a workforce constraint that no technology can fully solve: the expertise physician advisers need takes years to build and cannot be compressed.
“The governance is still being written, and it has to catch up with what the technology can already do,” Dr. Greer said. “And you can’t hire your way out of all your problems quickly as a result.”
The version of the next few years she wants to see is one where shared technology standards and transparency replace two sides looking at the same patient through different lenses.
“If technology closes that gap instead of widening it, then that’s the version of the next few years I’d want to see,” she said.
