As artificial intelligence (AI) becomes embedded into administrative workflows, healthcare organizations must adapt to keep up with changing standards for how claims are documented, coded, submitted, and disputed. Providers and revenue cycle management vendors (RCMs) use AI to improve documentation, accelerate claims submission, reduce denials, and strengthen reimbursement strategies. Though many of these capabilities can create real value for the healthcare system by reducing errors and improving efficiency, they also introduce new challenges for payers.
As provider tools become faster and more sophisticated, health plans need equal visibility into whether changing billing patterns reflect appropriate reimbursement or expose emerging payment integrity risks.
AI is being applied across the entire provider billing lifecycle, beginning before the patient encounter and continuing through claims submission, denial management, and appeals. From identifying care gaps or documentation opportunities at the beginning of an encounter to transcribing conversations, medical record entries, or translating documented services into billing codes, AI is helping providers submit cleaner claims.
RCM vendors position AI as a way to reduce administrative overhead while improving revenue performance, aiding in higher cost savings, improved productivity, and coding error and denial reductions. But without proper response frameworks and AI governance, these tools can cause negative outcomes.
A recent Cotiviti analysis of inpatient claims billed with Medicare severity diagnoses related groups (MS-DRGs) between 2021-2025 found a subset of providers whose billing showed a sharp jump in complex DRGs, especially between 2024-2025. These claims include more complications or comorbidities (CCs) or major complications or comorbidities (MCCs) as secondary diagnoses that lead to increased reimbursement. The average length of stay for those inpatient admissions does not show a similar increase during a later time period that could correlate to sicker patients.
Figure 1. Growth in proportion of MS-DRGs with CCs or MCCs within a subset of over 200 high-volume hospitals or hospital systems.
In one example, data around sepsis diagnosis codes tells a compelling story: diagnoses of sepsis with major complications rose without a corresponding increase in treatments, and average length of stay for these patients decreased. If more patients were genuinely experiencing sepsis with major complications, data would reveal more aggressive treatment and longer hospital stays. The absence of this data suggests that increased coding intensity may not reflect a true change in patient acuity.
When providers shift aggressively towards AI-driven programs without a system of checks-and-balances, tools that help optimize operations may also capture diagnoses inappropriately, misrepresent services, create inaccuracies in the member record, or contribute to inflated reimbursement.
As AI-assisted billing becomes more common, plans may face new pressure across payment integrity, medical policy enforcement, claim review, and appeals operations. Watch out for these pain points to better manage your AI-driven workflows and help ensure accuracy across the payment lifecycle.
-
Scale: AI can help providers generate, refine, and resubmit claims with greater speed, which may increase the volume and complexity of cases requiring payer attention.
-
Signal detection: Plans need to distinguish legitimate improvements in documentation and coding accuracy from patterns that suggest upcoding, inappropriate diagnosis capture, or services being represented in ways that are not supported by the record.
-
Pace: Plans should consider whether current policies, edits, and reviews are keeping up with fast-changing provider strategies.
AI-assisted coding can introduce more nuanced, data-driven billing patterns, where denial management may become more sophisticated as providers use AI to support resubmissions, appeals, and dispute strategies. Health plans need solutions that provide earlier visibility and enable monitoring of emerging billing trends simultaneously. By surfacing high-risk claims and distinguishing support defensible decisions with combined prepay AI and human review, plans can improve efficiency without creating unnecessary provider abrasion.
How peers are managing emerging AI-driven challenges
Health plans are at different stages of readiness. Peer experience points to a common theme: plans recognize the need to respond, but many are still determining what the right operational model should look like. Action starts with understanding the data and looking for changes in provider billing patterns, then evaluating whether those changes align with appropriate payment strategies.
To manage and mitigate risk, plans should strengthen payment integrity programs with analytics that can monitor emerging trends and prioritize reviews. Consider new policies, detection tools, and AI-enabled approaches that complement expert human review.
The long-term goal of payer-provider alignment around appropriate payment remains unchanged, but in the near term, plans need the ability to close the gap created when providers adopt AI more quickly than payers. By combining better visibility, stronger governance, and targeted intervention, plans can respond to AI-driven billing changes while continuing to support accurate reimbursement and a more efficient healthcare system.
AI-driven provider billing and its impact on health plans
Watch the latest Payment Integrity Pulse webinar on demand as Cotiviti experts explain how AI is shaping provider billing patterns, affecting health plans, and how to mitigate emerging risks.