Innovaccer keeps humans in the loop with CaduceausHealth deal

Making progress with autonomous revenue cycle management

CaduceusHealth's founder and CEO, Jim Bonomo, said there is a major difference between good and great in the revenue cycle.

He said in the announcement it "comes down to knowing which payers push back on which codes, shifts in auth requirements, or which denials are worth fighting."

Bonomo's company has built that specialized knowledge base over the past 30 years, providing billing, claim tracking and denial resolution services for physician practices.

However, the age of AI is dawning in revenue cycle management, with more processes being shifted to AI agents.

Most healthcare organizations use some form of AI in the revenue cycle, according to a 2025 survey conducted by the Healthcare Financial Management Association. That percentage has likely grown from the 63% of survey respondents since then, as about two-thirds also said that AI and automation will have the most significant impact on denials and underpayments. Nearly three-quarters also said AI will have the biggest impact on prior authorizations.

Revenue cycle technology vendors are leaning into this investment area, adding more AI capabilities to their platforms to satisfy providers' appetite.

Top vendors are even looking to make the revenue cycle autonomous using generative and agentic AI to completely take over certain tasks, such as front-end insurance verification, ambient medical coding and payment posting.

Innovaccer is one of those vendors, using its AI healthcare intelligence and data infrastructure to power its AI-native Flow suite. But the "goal isn't to remove humans," Shashank told RevCycle Management.

"It's to remove the administrative assembly work that consumes them," he explained.

Instead, AI can manage data assembly automatically, so when a case reaches a human, all the data they need is already there, he continued. Then, more cases get resolved, and staff spend less time on work that shouldn't require such a heavy manual lift.

Using AI to augment RCM expertise

Innovaccer has leaned heavily on the human-in-the-loop concept, which involves humans actively participating in a system's workflow to verify and correct AI and machine learning models.

In a report released earlier this year, Innovaccer outlined its vision for autonomous healthcare administration. While it relies heavily on AI agents for high-volume administrative work, human oversight remains critical at this highest level.

Most revenue cycle management vendors recognize they need human oversight, even as they drive their AI agents to operate more in a closed-loop fashion. This has left some industry experts questioning whether the revenue cycle can truly be autonomous.

But Innovaccer's acquisition of CaduceusHealth underscores how such a model could operate.

CaduceusHealth's certified revenue cycle management specialists will be embedded into Innovaccer's platform, according to Shashank, providing an "expert judgment layer."

"The human reviews, decides, and owns the outcome. The AI handles the repeatable majority; humans stay accountable for every case that requires judgment," he continued.

With the revenue cycle management experts, the vendor is also encoding decades of real-world experience into its AI models. This should train the AI models to be more accurate and operate more independently.

"That's the difference between AI that looks good on a demo and AI that actually improves net collections," said Shashank.

The deal also allows CaduceusHealth's experts to work at a scale "no managed services organization could reach on its own," Bonomo said. And for their clients, it provides a way to outsource revenue cycle management while still leveraging AI.

For example, Englewood Hospital and Medical Center grew from 50 providers and $30 million to more than 700 providers and $240 million in managed revenue using CaduceusHealth's services.

Innovaccer's acquisition of the company "means the next generation of providers won't have to choose between operational expertise and AI -- they'll have both," CFO Tony Orlando said in the announcement.

Jacqueline LaPointe is an Executive Editor at Xtelligent Healthcare Media, covering revenue cycle management, healthcare payers, health policy, and health IT since 2016.