CMS Taps AI to Combat Healthcare Fraud in New Federal Pilot

BOSTON, MA – March 23, 2026 – The Centers for Medicare & Medicaid Services (CMS) is launching a new offensive against healthcare fraud, waste, and abuse, turning to advanced artificial intelligence to modernize one of its most critical oversight functions. In a significant move, healthcare AI firm Basys.ai has partnered with ePathUSA, the prime contractor for a new CMS pilot program aimed at revolutionizing medical record review.

This collaboration, part of the CMS Artificial Intelligence Medical Record Review (AI MRR) pilot, will deploy specialized AI to help human reviewers analyze complex medical documents more efficiently and accurately. The initiative represents a major step in the federal government's push to leverage cutting-edge technology to protect taxpayer dollars and ensure the integrity of its vast healthcare programs.

A New Front in the War on Healthcare Fraud

Fraud, waste, and abuse (FWA) are persistent and costly problems within the U.S. healthcare system, siphoning billions of dollars annually from federal programs like Medicare and Medicaid. For years, the primary defense has been manual medical record review—a time-consuming, labor-intensive process where human auditors meticulously comb through documents to verify compliance and identify improper payments.

This traditional method, however, struggles to keep pace with the sheer volume of claims. Industry estimates suggest that 5-10% of all healthcare claims may involve some form of FWA, much of which is missed by manual processes that are better suited to finding errors after the fact than preventing them. The result is an administrative bottleneck that creates delays and significant burdens for both payers and providers, while allowing billions in improper payments to slip through the cracks.

"Medical record review has historically relied on time-intensive manual processes," noted Arpan Saxena, CIO of Basys.ai. "Applying domain-specific AI models and structured reasoning frameworks can help reviewers surface relevant clinical information more efficiently while maintaining the transparency required in regulated healthcare environments."

Deploying 'Agentic AI' for Program Integrity

At the heart of the new pilot is Basys.ai's 'agentic AI' platform. Unlike general-purpose AI, this technology is purpose-built for the complexities of healthcare administration. It is designed to act as a sophisticated assistant for human reviewers, capable of interpreting unstructured medical notes, understanding complex insurance policies, and cross-referencing information to flag potential discrepancies that might indicate FWA.

"This collaboration represents an important opportunity to apply agentic AI to one of the most complex operational challenges in healthcare," said Amber Nigam, CEO of Basys.ai. "Program integrity requires systems that can interpret complex medical documentation while maintaining accuracy, transparency, and alignment with regulatory frameworks."

Crucially, the system is designed for full transparency, avoiding the "black box" problem that plagues many AI systems. Every conclusion the AI reaches is auditable and traceable back to specific evidence within the medical record, a non-negotiable feature for operating within a regulated government environment. This focus on explainability has earned the Boston-based company significant recognition. Basys.ai was a top 10 finalist in the CMS "Crushing Fraud, Waste, and Abuse" Chili Cook-Off competition and has been featured by industry authorities like Gartner and the Harvard Business Review for its work on deploying reliable AI in regulated settings.

The Human-AI Partnership in a Regulated Space

The CMS pilot is not about replacing human experts but empowering them. The goal is to create a human-AI partnership where technology handles the painstaking work of data extraction and initial analysis, freeing up human reviewers to apply their critical judgment to the most complex cases. This collaborative approach is essential for navigating the ethical and regulatory minefield of healthcare AI.

CMS itself has established clear guardrails for AI use, emphasizing that technology must support, not supplant, human decision-making and oversight. Concerns over algorithmic bias, data privacy, and accountability are paramount. By designing its AI to augment human workflows and maintain a clear audit trail, Basys.ai aims to build the trust necessary for wider adoption.

"As the prime contractor for the AI Medical Record Review pilot, ePathUSA is focused on bringing together operational expertise and advanced AI capabilities to support CMS's program integrity mission," said Hari Nallure, President of ePathUSA. "Our collaboration with Basys.ai allows us to combine innovative AI technologies with practical implementation experience to evaluate how AI can improve the efficiency, accuracy, and transparency of medical record review."

A Strategic Alliance for a Federal Mandate

The partnership brings together a seasoned government contractor and a nimble AI innovator. ePathUSA, an Iowa-based technology services company, holds the prime contract with CMS, a two-year, $4.3 million agreement to test the viability of AI technologies in medical records. With a long history of supporting federal agencies like CMS and HHS on data engineering and systems modernization, ePathUSA provides the operational expertise and government contracting experience necessary to manage the pilot.

Basys.ai contributes the specialized, FWA-tested AI infrastructure. This synergy of a prime contractor with deep federal experience and a subcontractor with niche technological expertise is a common model for driving innovation within the government.

The success of this pilot could have far-reaching implications, potentially setting a new standard for how program integrity is managed across all federal healthcare programs. By demonstrating that AI can be deployed in a transparent, reliable, and collaborative manner, this initiative could pave the way for broader adoption of technology to make public healthcare more efficient and secure.

"Healthcare is entering a phase where AI will increasingly support complex operational decisions," Nigam added. "Our objective is to build agentic AI systems that healthcare institutions and regulators can trust - systems that improve efficiency while maintaining the transparency and accountability required in public healthcare programs."