Agency Leader Says AI Is Helping Resource-Strained Workforce Identify More Fraud

Why It Matters

AI amplifies CMS’s ability to detect and recoup fraudulent payments, directly reducing healthcare costs and improving the efficiency of a resource‑constrained agency.

Key Takeaways

• CMS CPI saved $26.3 B in FY2024, $14.6 ROI per dollar.
• AI assists 500 staff in reviewing 4‑5 million daily claims.
• $1 B budget faces $100 B fraud estimate, AI narrows gap.
• Pilot programs will suggest penalties, accelerating fraud recovery.
• Federal AI use cases rose to 3,611 in 2025, doubling 2024.

Pulse Analysis

Healthcare fraud remains a persistent drain on public finances, with estimates ranging from $100 billion to three times that amount annually. The Centers for Medicare & Medicaid Services (CMS) faces the dual challenge of a limited $1 billion investigative budget and an ever‑growing volume of claims. By integrating AI into its Center for Program Integrity, CMS can automatically flag anomalous patterns across millions of transactions, allowing auditors to focus on the most suspicious cases. This targeted approach not only maximizes the impact of a modest workforce but also generates measurable financial returns, as evidenced by the $26.3 billion saved in fiscal 2024.

The AI deployment at CPI goes beyond simple anomaly detection. Advanced models analyze claim histories, provider behavior, and cross‑referenced data sets to assign risk scores, effectively triaging work for investigators. The next phase introduces prescriptive analytics that suggest appropriate penalties, shortening the decision‑making cycle and increasing the speed of fund recovery. Compared with other federal agencies, CMS’s ROI of $14.6 per dollar invested outpaces many public‑sector AI pilots, highlighting the sector‑specific value of machine‑learning in high‑volume, high‑stakes environments.

CMS’s success reflects a broader trend of AI adoption across the federal landscape, with reported use cases more than doubling from 2024 to 2025. While agencies like the Department of Veterans Affairs are leveraging similar tools to expedite benefits processing, they also face scrutiny over error rates and fairness. For CMS, the challenge will be balancing rapid fraud detection with due‑process safeguards, ensuring that AI recommendations are transparent and contestable. As the technology matures, policymakers are likely to consider expanded authority and funding to fully harness AI’s potential in safeguarding taxpayer‑funded health programs.