Medicare Opens the Door to AI Reimbursement: September Deadline Could Reshape Clinical Payment Models

The Centers for Medicare & Medicaid Services has launched a policy inquiry that could fundamentally alter how artificial intelligence is valued and reimbursed in American healthcare. Buried within its Calendar Year 2027 Physician Fee Schedule Proposed Rule is a Request for Information asking a deceptively simple question: "What are the payment implications of including technology in primary care?" The comment deadline of September 14, 2026, represents a narrow window for physicians, health systems, and AI companies to influence how Medicare will pay for clinical services that increasingly rely on algorithmic decision support.

As AI moves from administrative back-office functions toward clinical decision-making—analyzing imaging studies, recommending treatment protocols, predicting patient deterioration—CMS is signaling that existing payment structures may no longer adequately capture the value or cost of these services. The agency's questions arrive at a moment when the healthcare industry lacks consensus on how to classify, govern, and compensate AI-enabled care.

The Classification Challenge: Assistive, Augmentative, or Autonomous?

The American Medical Association has developed CPT Appendix S, a taxonomy that classifies AI-enabled medical services into three categories: assistive (supporting administrative tasks), augmentative (enhancing physician decision-making), and autonomous (making clinical decisions independently). This framework, updated June 8, 2026, represents an attempt to bring conceptual order to a rapidly evolving technological landscape.

The distinctions matter enormously for payment policy. An AI system that analyzes thousands of previous total hip replacement procedures to identify patterns for implant selection, sizing, and positioning decisions might be classified as augmentative—enhancing surgeon judgment—or potentially autonomous if it directly recommends specific clinical actions. How CMS chooses to value these different categories will determine whether AI integration is financially viable for healthcare providers and whether reimbursement incentivizes applications that genuinely improve clinical outcomes or simply automate existing processes.

Beyond Coding: Questioning the Payment Architecture Itself

CMS's inquiry extends beyond AI-specific questions to challenge fundamental assumptions about how medical services are defined and valued. The agency is questioning its reliance on the AMA's CPT coding system and the AMA/Specialty Society Relative Value Scale Update Committee (RUC) for determining physician service valuation, according to the proposed rule (91 Fed. Reg. 43842, 43952–56).

This broader questioning suggests CMS recognizes that AI integration may require more than incremental adjustments to existing payment models. If AI systems fundamentally change clinical workflow—reducing time required for certain tasks while increasing complexity of others, shifting work from specialists to generalists, or enabling remote delivery of services previously requiring in-person visits—the entire architecture of fee-for-service payment may need reconsideration.

The Coalition for Health AI is developing governance frameworks to address these challenges, while Epstein Becker Green has launched a Digital Health Roundtable series to facilitate industry dialogue on these policy questions.

What's at Stake for Stakeholders

For healthcare providers, CMS's eventual payment decisions will determine whether AI-enabled services are financially sustainable, what evidence standards must be met to justify reimbursement, and whether payment models reward improved clinical outcomes or simply reduced labor costs. For AI companies, Medicare payment policy will directly influence market viability and product development priorities. For patients, the payment framework will shape which AI applications are deployed in clinical settings and whether AI integration genuinely enhances diagnostic accuracy and clinical judgment or adds technological complexity without meaningful benefit.

The September 14 deadline is approaching rapidly. Stakeholders who fail to comment may find themselves operating under payment policies developed without their input—policies that will shape healthcare AI deployment for years to come.

Sources

1. When AI Becomes Part of Medicine: Why CMS's AI Questions Matter Now — Epstein Becker Green