AI Adoption in denials management lags as other RCM uses expand

Denials management, one of the toughest tasks in hospital revenue cycle management (RCM), deserves more attention from AI experts, but RCM execs are not diving in with AI solutions for multiple reasons.

About one in five healthcare providers apply AI to denials management, a 2025 Bain & Co. survey found. Instead, providers have turned more of their AI investments to documentation support such as ambient listening (64%); clinical documentation improvement and compliance assurance for payer interactions (43%); and medical coding (30%).

What has held leaders back from making bigger AI moves in denials management?

Recent research, as well as interviews with healthcare revenue cycle leaders, point to the following factors.

1. There is still a lack of trust in what AI can achieve in the revenue cycle. Hesitance to deploy AI stems in part from unproven accuracy and skepticism that AI solutions will understand payer-specific rules, according to a 2025 Experian Health survey . One way that Nebraska Medicine builds trust in AI revenue cycle solutions is by taking its time with AI implementation and ensuring that staff are comfortable with its tools. “Make sure staff feel included in the validation process,” said Sheila Augustine, FHFMA, MHA, director of patient financial services for Nebraska Medicine in Omaha.
2. Denials management didn’t rise to the top of leaders’ AI wish lists until recently. Initially, priority applications of AI in the revenue cycle centered on other areas. Meanwhile, the Experian Health survey points to the potential for AI to derive value for insurance eligibility and benefits verification, patient scheduling/access and patient registration/data collection.
   
   At Rush University System for Health in Chicago, leaders rely on data to identify pain points that could be addressed with AI. Then, they examine whether the AI tools already available to them via Epic could be used to address those pain points or whether a more specialized tool is needed.
   
   “It’s a revolutionary time for AI innovation in revenue cycle, but when it comes to selecting use cases, I always equate it to the analogy, ‘You don’t go to a grocery store hungry,’” said Blake Evans, CHFP, system vice president for revenue cycle, Rush University System for Health. “You have to ensure you’re using data to determine the problem you’re looking to solve and the types of tools that will get you there.”
3. Other RCM tech implementations have captured time and resources. For example, at UMC Health System in Lubbock, Texas, intense focus on an EHR system implementation set for this year means leaders haven’t made strides in AI implementation — yet.
   
   “We are implementing Epic in early 2026, and we have denials mitigation in the forefront of our minds as we design workflows,” said Karen Veselsky, vice president of revenue cycle, UMC Health. That doesn’t mean that the revenue cycle team hasn’t begun to explore AI. One early use for the health system is drafting appeal letters with AI, Veselsky said.

Find more examples of AI use in RCM in hfm magazine’s story, “Battle of the bots intensifies over denials.”