Table of Contents
For years, Medicare has been wrestling with how to pay for artificial intelligence and other software-based clinical tools. CMS is good at calculating the costs of physical items, from a cotton swab to the wear and tear on a CT scanner, but an algorithm to predict cardiac risk from a CT scan, or an AI-based map to visualize prostate cancer’s spread, has been a much harder fit. Now the agency is proposing Software as a Medical Service, a new interim Payment category designed to close that gap.
Why Software as a Medical Service Is Needed
This month, in its proposed rules for hospital outpatient payments and physician fees for 2027, CMS has signaled that it’s ready to build a more consistent payment structure for clinical software and AI that factors in their impact on patient outcomes. It’s starting, as an interim step just for 2027, by proposing a practical change to the way it labels and pays for several clinical software and AI services.
A Renamed Category
CMS is proposing to rename what it previously called “Software as a Service” to Software as a Medical Service, stating that the SaaS label risks confusion with generic cloud-computing terminology used outside healthcare. The agency said it has been evaluating how to develop a comprehensive and consistent approach to SaMS payment for several years, given “the novel and evolving nature of these technologies.”
How Software as a Medical Service Would Be Identified
A new status indicator would flag these products for the first time. CMS is proposing to create status indicator “O1” and designate 36 Healthcare Common Procedure Coding System codes as SaMS. Of those, 21 would move out of standard clinical Ambulatory Payment Classification groups and into New Technology APCs, the payment track CMS typically reserves for new procedures that do not yet have enough claims data for a permanent home.
What Types of AI Tools Are Covered
The proposal covers a range of AI-assisted diagnostics. Examples listed in the rule include AI analysis of retina images for disease detection, echocardiogram-based heart failure detection, coronary blood flow estimates derived from CT angiography, CT-based bone fracture risk scoring, and eye-movement-based concussion assessment, giving a sense of just how broad the category of Software as a Medical Service tools is expected to be.
A Parallel Shift for Lab-Based Algorithms
A complementary shift moves 10 lab-data algorithms from the Clinical Laboratory Fee Schedule to contractor-priced payments under the Physician Fee Schedule. New codes describing the same type of lab-data algorithm would also be contractor-priced under the PFS beginning in 2027, if the proposal is finalized. This matters because the lab fee schedule operates under different pricing mechanics, cost-sharing rules, and budget-neutrality requirements than the physician fee schedule.
Why CMS Wants This Shift
CMS cited limited transparency into proprietary algorithm costs and potential program-integrity concerns as reasons to reconsider that route for algorithm-only laboratory services, aligning its approach across both the outpatient and physician fee schedule rules.
The Existing Gaps Software as a Medical Service Aims to Fix
A major barrier to the adoption of clinical AI technology in healthcare has been the lack of a clear Medicare benefit category. As of January 2026, there were 26 CPT codes for clinical AI solutions, but only three had received permanent Category I CPT codes: FFR-CT, a service analyzing CT scan data to predict cardiovascular risk; and a tool using augmented imaging to detect diabetic retinopathy. All other AI solutions with a designated CPT code remained Category III, temporary codes primarily used to collect data that may lack an associated payment rate and do not guarantee reimbursement.
Inconsistent Pricing Has Been a Persistent Problem
For many AI tools billed as physician services, CMS has relied on Medicare Administrative Contractors to set payment rates on a case-by-case basis, a process known as carrier pricing, which creates significant variation in how and where AI tools are reimbursed across the country. The Software as a Medical Service proposal represents an attempt to reduce that inconsistency with a more standardized framework.
What Hospital and Billing Teams Should Do Now
The immediate practical impact falls on hospital IT, informatics, compliance, and billing teams. Organizations offering genomic interpretation or similar algorithmic services should determine which current CLFS codes might move to PFS contractor pricing and project the revenue impact under both scenarios ahead of the proposed 2027 effective date.
This Is Still a Proposal, Not Final Policy
CMS will consider public comments and issue final rules later in 2026, likely adjusting the framework before the 2027 effective date takes hold. Organizations affected by the Software as a Medical Service proposal should treat the current framework as a planning exercise rather than a settled outcome, and consider submitting data-driven comments to the CMS-1801-P physician fee schedule docket to help shape the final structure.
What This Means for the Future of AI Reimbursement
The Software as a Medical Service proposal, paired with the AMA’s own consideration of a new Clinically Meaningful Algorithmic Analyses coding classification discussed at a December 2025 meeting, suggests that both CMS and organized medicine recognize the current coding and payment framework for clinical AI is inadequate for the pace of technology adoption. As these parallel efforts develop, the coming months could bring meaningful clarity to how AI diagnostic tools get priced and reimbursed under Medicare going forward, a shift with implications well beyond the 36 codes initially named in this proposal.
For more healthcare industry updates, insights and news, visit DistilINFO. Click here to subscribe to stay informed.
