Health System AI: A New Frontier for Payment Integrity

Rajeev Ronanki, M.S.

Artificial intelligence (AI) is rapidly transforming hospital operations, from diagnostics and documentation to revenue cycle management. As these tools grow more powerful and pervasive, they are reshaping how care is delivered, documented, and billed. For health plan leaders, this new era introduces both enormous potential and significant risk. AI may streamline care, reduce waste and optimize workflows. But left unchecked, it could also be used to inflate billing, obscure audit trails, and push utilization into more expensive settings.

AI use cases

Hospitals and health systems are deploying AI across clinical, operational, and financial domains. On the clinical side, AI is being used to interpret radiology images, support diagnoses, and personalize treatment plans. Operationally, AI-powered scribes help reduce documentation burdens, virtual assistants triage patient queries, and predictive models manage staffing and supply chains.

Most importantly for payment integrity, AI is rapidly expanding into the revenue cycle. Natural language processing (NLP) tools are generating visit notes from audio recordings, which are then translated into billing codes using generative AI. Other tools recommend the highest reimbursable diagnosis based on the chart. Automated systems are being used to accelerate prior authorizations, optimize place-of-service selection, and drive more aggressive appeals workflows. According to the Medical Group Management Association (MGMA), nearly half of large provider organizations expanded their AI footprint in 2024 alone, with many identifying revenue enhancement as a key goal.

Potential benefits

AI has the potential to deliver real and measurable improvements across both patient care and health plan operations when deployed appropriately. AI-enhanced diagnostics can catch early signs of disease, allowing for more timely and less costly interventions. Machine learning models can personalize treatments based on patient history and predicted outcomes, improving care quality and reducing complications. Tools like ambient documentation help ease the administrative burden on physicians, improving record accuracy while also reducing burnout. Predictive analytics further strengthen care delivery by identifying high-risk patients and potential care gaps before they escalate into costly events.

For health plans, these capabilities translate into improved outcomes and reduced utilization. AI that supports value-based care models can help close care gaps and ensure timely, appropriate services, leading to fewer unnecessary tests, better medication adherence, and faster claims adjudication.

In short, when used as intended, AI can align incentives across health plans and providers by driving down waste and improving care quality. As these tools proliferate, they will fundamentally reshape the volume, complexity, and nature of claims and appeals — often in ways that are difficult to detect with traditional rules-based systems.

Risks and red flags

The same AI tools that can improve care and efficiency also carry the potential for abuse. Some hospitals are already training AI to optimize billing codes that maximize reimbursement even when clinical justification is thin. AI-generated documentation may include diagnoses or symptoms that were never discussed during the clinical encounter. Algorithms might suggest longer hospital stays or assign more complex codes to routine procedures. In these cases, AI becomes a multiplier for overbilling rather than a tool for care optimization.

One recent example underscores this risk. In late 2024, a major Colorado health system agreed to pay $23 million to resolve allegations of using automated rules to upcode emergency department visits. Federal prosecutors cited concerns about systemic misuse of documentation tools to generate inflated bills. Similar investigations have since been opened into other organizations allegedly using AI to surface questionable diagnoses or deny coverage based on opaque algorithms.

Another emerging concern is the use of AI to generate high volumes of appeals. These tools can automatically populate templated arguments, source related documentation and submit requests at scale. Although this can streamline appropriate appeals, it can also overwhelm health plan operations if misused, outpacing internal staffing, delaying resolution and driving up administrative costs.

When used tactically, mass appeals can become a blunt instrument to force claim reconsiderations or slow denials. This burdens payment integrity teams, complicates claim workflows and inflates operational overhead. Without constraints, such tactics could result in higher claim volumes, more appeals and rising costs for both employers and members. Health risk being overwhelmed by AI-inflated claims that appear legitimate on the surface. If these shifts in billing behavior go undetected, the resulting regulatory backlash could be substantial.

Early trends

Health plans are already seeing early signs of change. Coding patterns are shifting. Claims are arriving faster. Documentation appears more polished but may not accurately reflect the patient encounter. These changes are subtle, but they point to a rapidly evolving landscape where AI is increasingly shaping how care is documented, coded and reimbursed.

Traditional payment integrity approaches — built on manual chart reviews and static rule sets — will not be enough. To stay ahead, health plans must be able to distinguish between AI-driven efficiencies and tactics that exploit the system.

Failing to act could result in increased costs, rising provider abrasion and heightened legal risk. Without a proactive strategy, health plans may find themselves losing ground in both cost containment and provider trust.

Evolving strategies

Leading payment integrity vendors are already helping health plans prepare for this new AI-driven landscape. This begins with building AI tools that can detect new behaviors as they emerge. Machine learning models can identify outlier billing patterns based on historical trends, peer benchmarks, and clinical plausibility. Pre-payment analytics can flag suspicious claims before dollars go out the door, while post-payment audits are strengthened by NLP tools that detect over-documentation or misaligned diagnoses.

But more importantly, the future of payment integrity lies in collaboration. Health plans must engage providers in transparent conversations about how AI is being used and monitored. New review processes should be designed not to penalize innovation but to uphold integrity.

The goal is not to slow progress but to guide it. Companies are investing in capabilities to help health plans walk this tightrope. By combining explainable AI with human oversight, these solutions enable faster reviews, smarter targeting and less friction. Payment accuracy platforms are designed to learn in real time, adapt to new patterns, and elevate only the right claims for review, allowing compliant claims to flow through seamlessly.

As AI adoption across health systems is accelerating, the future of healthcare is being actively rewritten. For health plans, this presents both an opportunity and a challenge. The opportunity lies in aligning with providers to deliver better care at a lower cost. The challenge lies in preventing misuse that undermines trust and drives unnecessary spending.

Payment integrity leaders must rise to meet this moment by modernizing their tools, sharpening oversight and strengthening partnerships. With the right strategy, AI can enhance — not compromise — the integrity of care. And with trusted partners, health plans can build a payment integrity function that is future-ready, data-driven and built to protect value across the ecosystem.

Rajeev Ronanki, M.S., is CEO of Lyric, a payment integrity company.