AI Targets Revenue Leakage as Healthcare Systems Rework Revenue Cycle Operations – PwC – June 3, 2026

By Staff Reports - June 3rd, 2026

Healthcare providers are reworking revenue cycle operations as administrative complexity and reimbursement gaps continue to erode margins. Billing, coding, and payer interactions remain a major cost center, while denials, underpayments, and process breakdowns limit cash realization and increase cost-to-collect across the revenue cycle. Adding insult to injury, fragmentation across systems, workflows, and data sources compounds the problem, making it difficult to trace where revenue is lost or delayed. 

In a recent BizTechReports executive vidcast interview, Jacob Shurbet, a principal at PwC, described the operational reality behind that pressure. “Your revenue cycle is a continuum of operations, and you have to get everything right for it to actually work,” Shurbet said. 

Small breakdowns at any point in that continuum, from eligibility verification, coding accuracy, documentation completeness, to payer adjudication, will likely propagate downstream, increasing rework, delaying payment, and ultimately suppressing realized revenue. 

From Cost Center to Strategic Lever

Revenue cycle performance sits at the center of margin management for healthcare organizations. Cost reduction remains a primary objective, but it no longer defines the overall opportunity.  

“Most people quickly go to reduce cost…but there’s also an upside and an additional yield component that senior leaders should consider,” Shurbet said. 

That upside takes the form of revenue already earned but not yet collected. It is in this context that AI can help organizations identify underpayments at scale, surface contract discrepancies, and recover dollars that would otherwise remain buried in transactional noise. Even small percentage improvements can make a big difference when applied across millions of claims, explained Shurbet.   

This is because AI exposes revenue leakage embedded in billing, coding, and payer interactions. It can catch underpayments, missed charges, and contract misalignments that often accumulate across transactions. In this scenario, additional revenue capture comes from executing a more precise and intelligent lifecycle claim strategy rather than increasing patient volume.   

“Payment integrity becomes a measurable outcome in which agentic systems monitor claims at a granular level and flag discrepancies in near real time,” said Shurbet. 

That requirement for precision brings into focus the structural barrier that has limited prior improvement efforts. 

Fragmentation Remains the Central Constraint

Because AI adoption has outpaced organizational readiness, many health systems are experimenting without establishing the structures required to scale results. 

Citing data from the Healthcare Financial Management Association, Shurbet noted that most organizations are using AI in some capacity while far fewer have mature governance in place to extract value from those investments.  

“Everybody wants to use AI… but there’s still a lot of fragmentation of what and how things are done. This gets in the way of achieving a coordinated outcome,” he said. 

For healthcare leaders, this fragmentation translates into stalled initiatives, inconsistent outcomes, and an inability to connect AI investments to measurable financial improvement. While point solutions based on AI can generate localized gains, they will likely fail to move enterprise-level metrics, such as days-in-accounts-receivable or net-collection-rates. 

This is because core systems generate large volumes of data but operate independently which results in electronic medical records, billing platforms, and clearinghouses not being shared in a unified data model. 

“Data is the ground zero for AI,” Shurbet said. 

Integrating this environment will determine whether AI operates as a point solution or as a system-wide capability. 

Resolving fragmentation requires a shift in how automation itself is defined. 

The Shift from Automation to Orchestration

When integration happens, AI can change the unit of work from task execution to outcome delivery. In practice, that means resolving a denied claim without human intervention or confirming that each service rendered is coded, billed, and reimbursed correctly the first time. 

“Historically, automation was about bits and pieces of a process… we’re saying, what is the outcome you’re trying to get to? And how can we leverage AI to achieve strategic results?” Shurbet said.  

The question is easier posed than answered because revenue cycle workflows contain constant variation. Each claim reflects different clinical conditions, payer rules, and contract terms. Properly implemented, AI agents can assess context, determine next actions, and coordinate across systems. The process adapts to the claim rather than forcing the claim through a fixed workflow. 

That shift from task execution to outcome accountability changes how work is distributed across both technology and people. 

Workforce Implications: From Execution to Oversight

Labor constraints already limit the ability to manage transaction volume. AI can change how that work is distributed, with agents handling routine claim follow-ups, eligibility checks, and documentation gathering while humans focus on exceptions and escalation paths. 

“We’re really looking at using AI to empower our staff to do better, to do things faster,” Shurbet said. 

Initial deployment focuses on augmentation in which staff complete the same tasks with greater speed and accuracy, supported by systems that surface next-best actions, required documentation, and payer-specific rules in real time. Time spent on research and manual coordination declines, allowing more accounts to be processed with the same resources. 

Work changes as AI assumes more of the transactional load.  

If phase one centers on augmentation, then phase two shifts responsibility toward managing exceptions and overseeing agent-driven workflows.  

“In phase two, the nature of their work is going to change… we’re going to have to rethink the role staff play,” he said. 

Phase three introduces a model where staff configure, train, and refine AI agents themselves. In this transition, human effort shifts toward exception handling, judgment, and negotiation. 

That evolution places new demands on the underlying technology environment. 

Infrastructure as the Enabler

Execution depends on consolidating fragmented data into an environment where AI can operate continuously, monitoring claims in flight, triggering interventions when discrepancies occur, and updating workflows in real time based on payer responses. 

PwC and Amazon Web Services (AWS) recently announced a joint initiative to help healthcare organizations modernize revenue cycle operations using AI and cloud-based data infrastructure. 

“We felt like AWS’s presence in the healthcare space… was a good foundation that we could build upon,” Shurbet said. 

Specifically, PwC brings domain expertise in revenue cycle operations, process redesign, and governance frameworks required to operationalize AI at scale. AWS provides the cloud infrastructure, data architecture, and AI services needed to aggregate, process, and act on large volumes of healthcare data securely. Together, the integration enables a unified operating model that connects fragmented systems, allowing AI to drive continuous, end-to-end improvement across the revenue cycle. 

As a result, existing systems remain in place, but an orchestration layer aggregates data, feeds AI models, and distributes outputs back into operational systems. 

“We’re thinking about how you take all of that data, build an AI orchestration layer… and feed your agents on what needs to happen,” he said. 

Early Signals of What Comes Next

Initial deployments demonstrate how quickly targeted capabilities can move into production. 

“In one case, it took about 10 weeks to build and deploy a conversational AI agent that was able to make outbound calls to payers,” Shurbet said. 

The system now initiates contacts, exchanges required information, and captures responses for integration into billing systems. 

“That just shows you the potential,” he said. 

These early successes are beginning to shift expectations around what can be automated and how quickly value can be realized. 

Leadership as the Deciding Factor

While the potential of strategic integration of AI and agentic systems is becoming more clear to the industry, adoption remains uneven – despite widespread experimentation. This is because some organizations deploy isolated use cases that never scale, while others struggle to move beyond pilot programs due to data constraints or unclear ownership. 

“Only 12% of CEOs say AI has delivered both cost and revenue benefits,” Shurbet said. 

This poses a challenge to both leaders and the rank-and-file, because technology does not align itself. Data, processes, and organizational ownership should be coordinated by staff to pursue a strategic vision that is well articulated by senior leadership. 

“It’s not only an AI problem, it’s a cultural problem… a data problem… an adoption problem,” he said. 

Cultural resistance, he added, often slows adoption when teams continue to rely on legacy workflows.  

“AI is not just an easy button,” Shurbet said. “You’ve got to have a good strategy, a good plan, a good data foundation…and then actually go after that.”