Fighting Healthcare Fraud with AI Arbitration Layers

Thank you to Mark Selcow for his collaboration in this piece 

Introduction

Fraud, Waste, and Abuse (FWA) is one of healthcare’s most expensive yet invisible challenges, consuming up to an estimated $300 billion annually (3-10% of total spending). This massive drain and wastage occurs through a web of misaligned intentions – inflated billing codes, phantom services, duplicate discounts – alongside systemic slippage and oversight. In a classic tragedy of the commons, honest stakeholders subsidize fraudulent actors through higher costs.

The core issue is fragmentation: disconnected data systems and opaque processes among payers, providers, and regulators create exploitable gaps. And the time is now: Healthcare organizations face rising margin pressure, while regulators increasingly demand proactive, real-time fraud prevention rather than slow, expensive post-payment recovery.

The scale and urgency of the issue makes it the perfect market opportunity for AI-native software solutions to tackle FWA. We believe intelligent, proactive monitoring is the first step toward becoming neutral AI arbitration layers, pre-adjudicating claims and triggering corrective actions before payments occur.

The potential benefits span every stakeholder:

• Payers prevent losses upfront rather than pursuing expensive post-payment recovery
• Providers get reimbursed faster with reduced administrative burden
• Patients see lower premiums and out-of-pocket costs as fraudulent expenses stop inflating the overall cost pool.

Where We’re Looking: New Arbitration Layers in Healthcare

This is not a new problem. Over the past two decades, both public programs and private vendors have attempted to address FWA, yet most efforts have underdelivered. Legacy approaches which were anchored in rigid rules, retrospective audits, and manual review have proven brittle in the face of evolving fraud schemes and growing data complexity.

Established incumbents focused heavily on post-payment reconciliation, relying on rule-based detection and human-led auditing. While many are now shifting toward pre-pay analytics, their infrastructure and business models are not optimized for real-time prevention. Valuations for companies like Cotiviti (acq for ~$4.9B in 2018) on the payer side and Model N (~$1B) on the pharma side, reveal the sheer magnitude of the problem.

We believe there’s an opening for new AI native entrants. Platforms can serve as arbitration layers between payers and providers, with the necessary data infrastructure to link disparate data. In doing so, these systems can surface context-aware anomalies across unstructured and structured data before payment: reading physician notes, interpreting billing narratives, and detecting patterns across claims data. Here’s where we’re looking:

Intelligent Monitoring and Pre-Payment Reconciliation for Payers

The systemic drains on payer economics from FWA are particularly acute in traditional claims systems that process individual transactions in isolation. Payers lose billions annually to preventable leakage, especially in complex verticals. A new generation of AI-native platforms is flipping the model from reactive recovery to proactive prevention:

• Falcon Health builds AI-driven infrastructure for fraud and payment integrity across Medicare, Medicaid, and commercial lines. 
• Onos Health helps payers track and manage high-risk, under-structured verticals like behavioral health, where services are difficult to standardize and validate.
• Bluespine discover, recovers, and prevents medical overbilling for employer-sponsored plans.
• Claimsorted operates as a modern Third Party Administrator that integrates real-time FWA detection and intervention directly into claims flow.
• Lilac is building an AI-native data orchestration and agentic automation platform for health plans. In automating data aggregation, cleansing, and normalization, they provide key insights and execute workflow to improve payers’ Medicare Advantage (MA) STARS ratings.

These platforms shift the model from “pay-and-chase” to “score-and-prevent,” offering payers faster throughput, better accuracy, and measurable cost containment.

P&L and Compliance Agents for Providers

While payer-side fraud often dominates the narrative, the largest FWA losses still originate at the provider level through upcoding, phantom billing, medically unnecessary procedures, and documentation gaps. Fixing this requires confronting the root cause: providers operate with limited visibility into their real-time financial and regulatory exposure.

Most hospitals and health systems lack embedded tools to proactively catch errors or detect compliance issues at the point of documentation or claim submission. Instead, they face denial weeks later or federal audits years down the line.

We’re excited about a new category of P&L and Compliance agents for providers: AI-native tools that sit inside provider workflows and surface financial, compliance, and coding risks. These systems not only reduce audit exposure but act as internal FP&A/Finance Team co-pilots:

• Phare offers automated coding validation, and uses AI to perform audits that catch and correct errors before they impact providers’ bottom line with a pre-bill safety net.
• Translucent creates Profit and Loss (P&L) agents for FP&A teams, aiming to automate the entire stack of complex financial backend for large provider groups.

Dispute Resolution Automation

Payer-provider disputes cost the U.S. healthcare system billions of dollars annually, driven by opaque workflows, subjective interpretation of coverage rules, and labor-intensive manual review. The recent introduction of the No Surprises Act’s Independent Dispute Resolution (IDR) process has only intensified the volume and complexity of claim-level arbitration.

We believe there is an emerging category of AI-first data infrastructure that can scale adjudication by reading a huge corpus of regulatory information and supporting data, and automating much of the manual paperwork that bottlenecks arbitration processes.

• Radix encodes regulatory logic and applies real-time decision support to disputed claims serving as an auditable adjudication layer.
• In adjacent domains like medical-legal work, Codes Health is building full-stack automation for medical chronology review.

These tools can compress arbitration cycles from months to minutes, offering enormous leverage across payer operations, third-party administrators, and legal-medical review entities.

Pharma Revenue Integrity Platforms

Pharmaceutical pricing channels are riddled with opaque rebate structures and copay fraud. This is of high financial impact. Gross to net leakage routinely strip margin from gross drug sales, compressing EBITDA. Inaccurate rebate accruals and post-pay recoveries create earnings swings and reserve adjustments. Regulatory risk is high where 340B duplicate discounts, Medicaid best-price violations and false-claims exposure trigger fines and Department of Justice scrutiny.

Some examples of fraud are pharmacy-level switching where pharmacies change prescriptions from lower-cost drugs to higher-commission alternatives, or copay program abuse where uninsured patients create multiple fake identities to repeatedly access manufacturer copay assistance programs.

We’re looking for platforms that bring real-time verification, rebate auditability, pharmacy management and monitoring into pharma’s hands. These tools should help pharma navigate increasing regulatory pressure from the Inflation Reduction Act, and Centers for Medicare & Medicaid Services integrity audits. 

Conclusion

We think the opportunity to reduce FWA with AI-native software can have a huge economic impact through both potentially huge cost savings and increases in market efficiency, while also promoting fairness in healthcare by shifting resources back to patient care. 

If you’re tackling FWA with software, let’s talk! Please reach out to Nicole or Mark.