The AI arms race reaches revenue cycle management: How leaders are responding

As payers deploy AI at scale to manage and deny claims, it’s clear: Fighting AI-powered denials with manual processes is no longer viable. The only way for providers to keep pace is to match the sophistication of the payers’ approach with their own investments in advanced tools for claims management and denials response.

A recent MDaudit Benchmark Report reveals the extent to which the AI arms race in revenue cycle management — with both payers and providers leveraging AI to manage claims — is impacting healthcare providers’ bottom line. Today, most providers are fielding major increases in denials and underpayments. Hospital inpatient and outpatient claim denials rose by double digits in 2025: 14% for hospital outpatient services and 12% for inpatient care.^a

Also a top concern: a surge in payer audits, with a 20% rise in cases per customer, according to the report, pushing up the cost to collect. The average amount per claim also increased by 18%, pushing up the cost to collect and putting revenue cycle leaders under intense pressure to mitigate this risk.

“It’s time to boldly apply AI tools to surface the payer patterns driving the greatest financial exposure in the healthcare revenue cycle and respond with agility,” said Ritesh Ramesh, CEO, MDaudit. “The providers that are gaining ground in the AI arms race are the ones that are combining technological and human expertise to catch high-dollar risk faster and respond sooner.”

In this HFMA roundtable — featuring senior leaders from large academic medical centers, regional health systems, a rural hospital collaborative and a state hospital network — participants offered an insider’s view into ways they are navigating the AI arms race between payers and providers. They also shared insight on how to leverage the right tools and processes to combat denials, underpayments and takebacks.

We know payers are increasing their investments in AI for claims management, resulting in a doubling down of denials, increased payer audits and clawbacks. What impacts are you seeing in your organization?

Tanya Sanderson: Having spent several years on the vendor side working in the AI space, it has been interesting to return to the provider side and see the evolving payer landscape from this perspective again. As payers have expanded their use of AI and automation in claims management, we’re seeing increasingly sophisticated challenges, including downcoding, underpayments and greater audit activity.

While there has been considerable focus on reducing the administrative burden associated with prior authorization, some of that burden appears to be shifting downstream. Providers are managing a growing volume of post-payment audits and payer reviews, often without standardized processes or communication channels. Requests may arrive by email, spreadsheet, fax, or other methods, creating significant administrative complexity for our teams.

Jane Kennedy: We’re seeing more underpayments with and without line-item denials, but what is the true denial? It takes a lot more work from our clinical and non-clinical teams to reach the payer and figure out what the underpayment and/or denial really is.

Tammy Dumlao: The challenge is that we don’t have the staffing to fight all of the appeals we get. The payers tell us that we have to go through the appeals process, but each one of them has a different appeals process. We heavily depend on vendors to help us with our follow-up and our denials, but they don’t currently have that AI capacity to help us out, either. So that’s where it’s a challenge for us right now. Sometimes we go through the insurance’s contracting department to address variances in payment, but those discussions often go around and around in circles. We’ll call the payer representative directly, but then they disappear, or they change jobs, and then you don’t know who to contact. In the meantime, we’re either writing the claim off because it’s not worth the effort, or we’re holding onto it until we get to a larger dollar amount with multiple claims. That’s where efforts really escalate.

Trey Malone: Payers have a much simpler job than healthcare providers. I don’t mean to oversimplify this for the people who work on the payer side, but it is a simpler task to be a payer than it is to take care of buildings full of people with limited capital resources. You’re dealing with a much more concentrated, focused type of work as a payer than as a provider.

Has your organization adopted autonomous coding as part of your denials defense? If so, are you seeing the results you expected?

Bradley Becker: Where AI or other types of automation have been rolled out, we’ve seen the greatest initial value on the clinical side. Since our businesses provide healthcare, it makes sense that, as we’re prioritizing, automation is currently used for the clinical side. But then we also have organizations that are using AI and other automation for internal mid-cycle revenue cycle processes first. This is an area that obviously needs to be focused on more.

Kennedy: We are really trying to get ahead of denials and underpayments with AI and automation. We’re using our own analytics tools to track denial trends; ‘We have this payer and we’re seeing this spike, and it’s all similar, so where is it coming from?’ You almost must have AI to combat AI, because when the payer has AI, they’re getting quicker at issuing the denials.

As we’re adopting AI tools, we’re working on keeping humans in the loop and elevating the expertise of the staff that we have. Automation and AI aren’t replacing the teams we have. It’s about taking our current skill sets and upskilling them.

Many payers are automatically downcoding certain DRGs [diagnosis-related groups] and CPTs [Current Procedural Terminology codes]. This is not just a revenue risk, but a compliance risk. How are you identifying downcoding? How are you combating it?

Sanderson: One of the challenges with payer-driven downcoding is that when providers challenge the reduced payment and seek reimbursement consistent with the documented level of care, the claim may ultimately move from a partial payment to a full denial. That creates a difficult dynamic for providers: Accepting the downcoded payment may leave legitimate reimbursement on the table while challenging it can introduce additional financial and administrative risk.

Addressing these practices requires a more deliberate and coordinated approach to payer accountability. Organizations need strong mechanisms to identify patterns, validate that reimbursement aligns with the clinical record and contractual terms and establish appropriate escalation pathways when systemic issues emerge. Building that infrastructure and strengthening payer governance are increasingly important as these practices become more sophisticated.

Kennedy: We established a multidisciplinary review committee that brings all key stakeholders to the table, including utilization review, physician advisors, clinical documentation improvement [CDI], contracting, legal and more. Together, we review case examples to identify denial trends and understand what we see across payers. Often, these issues are not isolated to a single diagnosis-related group [DRG] but span multiple DRGs and clinical scenarios.

From the CDI and coding perspective, we also take the opportunity to reassess our internal standards and documentation practices. For example, our organization has established organizational criteria that align with physician documentation expectations. We evaluate whether our clinical and coding standards provide a defensible position and whether we have sufficient justification to challenge payer determinations. If that justification is sound, we then consider whether those expectations should be further reinforced through future contract language to better support our organization and reduce ongoing disputes.

Becker: One of the challenges is aggregating the data and telling the entire [denials] story, because things are happening and changing real time, and it is difficult to go back retrospectively. So how do we track this? You have to come up with an additional adjustment code. Some of our hospitals started to keep a log and then can go back and produce a report. Otherwise, the new denial information can be buried within other denial reasons, and you’re challenged to pull it out. Because most of our facilities are critical access hospitals, they’re not paid under the DRG system, so then there are line-item denials that aren’t accepted, and again, they’re buried in the percent of charges payment methodology. The CPT downcoding is occurring more and more on the outpatient side, where there is voluminous data. That’s where you have to somehow identify and record it so you can track those trends and leverage them during discussions with payers.

The current state of healthcare organizations reveals a strong synergy and collaboration between stakeholders along the revenue cycle continuum, resulting in positive outcomes. Where is your organization at in this transformation?

Sanderson: A key component of our transformation is bringing teams together across the revenue cycle and clinical continuum. Silos still exist, particularly between clinical and financial functions. As a nurse by background who has spent much of my career in healthcare finance and revenue cycle, I see an important opportunity to more intentionally integrate the clinical perspective into revenue cycle strategy.

Too often, clinicians become involved after a denial has occurred, particularly during the appeals process. The greater opportunity is to engage clinical partners further upstream, where their expertise can help identify risk, strengthen prevention strategies, and address issues before they result in denials or lost reimbursement. Building those partnerships across clinical, operational and revenue cycle functions is a critical part of creating a more proactive and sustainable approach to revenue integrity and financial performance.

Kennedy: My organization has a clinical revenue cycle team within our broader revenue cycle management structure. CDI, coding, denials management, utilization review and other clinical teams are all integrated into that space. What I appreciate most as a clinician is that it connects the patient-centered care we provide at the bedside with the work of our finance teams. Together, we recognize that every role impacts the patient experience and outcomes.

We’re a team from end to end. I don’t want to say the silos are gone, but we’re working to break those barriers down. We all have a seat at the table because every voice matters, and the diverse perspectives of our team strengthen our decision-making.

Conclusion

The theme of the 2026 HFMA Annual Conference, “Dare to Solve,” could be described as the motto today’s revenue cycle leaders are living by. The conversations among leaders participating in this roundtable made it clear: These leaders aren’t waiting for the perfect conditions or timing to bridge the payer-provider divide. They are solving challenges with the urgency the moment demands.

“For revenue cycle leaders, the dare in ‘Dare to Solve’ includes building the infrastructure to hold your ground in the AI arms race,” said Ramesh of MDaudit. “The leaders whose teams are operating as one entity, with shared data and goals, are the ones who feel best equipped to navigate this challenge. These are the teams that are turning one-off disputes into a coordinated payer strategy — and their organization’s financial health is better off for it.”

Panelists

BRADLEY BECKER,
FHFMA, MHA, is senior director of payer strategy at Mason General Health in Shelton, Wash., and The Rural Collaborative in Olympia, Wash.

TAMMY DUMLAO,
CRCR, EHRC, is director of revenue cycle for Hawaii Health Systems Corporation in Honolulu.

JANE KENNEDY,
BSN, MSW, is enterprise director of utilization review and inpatient appeals for WVU Medicine in Morgantown, W.V.

TREY MALONE
is regional director of financial operations for Ardent Health Services in Amarillo, Texas.

DANA FINNEGAN
is the senior director of market strategy for MDaudit in Waltham, Mass.

TANYA SANDERSON,
CHFP, MBA, MHA, is senior director of denials management at Stanford Health Care in Palo Alto, Calif

About MDaudit

MDaudit is an award-winning, AI-powered, continuous risk-monitoring platform and trusted revenue integrity partner to healthcare organizations nationwide. Working in the background, MDaudit delivers the insights organizations need to face the future with confidence — enabling teams to achieve more with less in a rapidly evolving environment. Learn more at www.mdaudit.com.

Footnotes

a. MDaudit, 2025 benchmark report: Key revenue integrity and compliance trends shaping healthcare, November 2025.