Jeff Francis, CFO at Nebraska Methodist Health System
Photo: Nebraska Methodist Health System
At Nebraska Methodist Health System, the revenue cycle operationally used to rely heavily on manual, repetitive work.
THE CHALLENGE
Staff spent significant time checking claim status, navigating payer portals and reacting to denials after the fact rather than preventing them up front. This reactive approach increased the cost to collect and slowed cash flow – especially painful in an environment of tight margins, workforce shortages, inflationary pressure and shifting payer mix.
From a leadership perspective, the magnitude of the problem was apparent but difficult to solve with traditional approaches, said Jeff Francis, CFO at the health system.
"Internal analysis showed that nearly 8% of total revenue was at risk annually due to documentation and coding issues, which for a billion-dollar organization translates into tens of millions of dollars," he explained. "That level of leakage directly limited our ability to reinvest in patient care, infrastructure and innovation.
"The core challenge was revenue leakage driven by incomplete or inaccurate documentation and coding across a high-volume, complex revenue cycle," he continued. "Clinical teams were delivering care appropriately, but the supporting documentation often failed to fully or consistently capture the patient story in a way that aligned with payer requirements. As a result, claims were delayed, underpaid or denied – putting a meaningful percentage of earned revenue at risk."
Many claims challenges originate upstream in the mid-cycle, where incomplete or imprecise documentation and coding create avoidable risk long before payers ever touch the claim, he added.
PROPOSAL
At the proposal stage, the promise was not automation for its own sake, but consistency and scale that manual processes could not achieve. For Nebraska Methodist Health System, AI-powered claims and coding vendor Akasa proposed addressing two distinct challenges within the revenue cycle: operational inefficiency in back-end claims work and accuracy risk earlier in the mid-cycle.
"On the back end, the proposal focused on using AI to automate high-volume, rules-based claim status activity after billing," Francis explained. "The intent was to reduce the amount of staff time spent manually checking payer portals and tracking routine claim updates, allowing revenue cycle teams to focus on exceptions, denials, and complex follow-up that require human judgment.
"Methodist is now entering its seventh year on claims updates and it would be a major dissatisfier if this function were taken away," he continued. "More recently, the vendor agreed to emphasize applying generative AI earlier in the revenue cycle to strengthen coding accuracy before claims are submitted. This approach was designed to ensure the full patient story was consistently and correctly translated into billable codes upfront, reducing downstream rework, denials and underpayments."
Critically, both use cases were built around a "human-in-the-loop" model rather than full automation, he added. AI would provide scale and consistency by surfacing patterns, risks and opportunities, while experienced revenue cycle staff at Nebraska Methodist retained final decision-making authority to ensure accuracy, compliance and revenue integrity.
MEETING THE CHALLENGE
The technology was integrated directly into Nebraska Methodist's existing revenue cycle workflows, allowing teams to use it within their daily operations rather than as a standalone tool. Different parts of the organization engaged with the technology in different ways, depending on the task.
In both cases, the AI augmented staff rather than replacing them. Integration with existing electronic health record and revenue cycle systems ensured adoption without major workflow disruption or additional IT burden, Francis said. The data-rich dashboards have also proved efficient in gauging performance and ROI, he added.
"On the back end, patient financial services and billing teams use AI to automate routine claim status work after claims have been submitted," he said. "The system continuously monitors large volumes of accounts, handling repetitive payer status checks in the background. This allows staff to step away from low-value portal work and focus instead on complex claims, denials, appeals and payer-specific follow-up that requires experience and judgment.
"We also use AI earlier in the mid-cycle to support pre-bill coding review," he continued. "Coding teams use the technology as a second set of eyes before claims are sent, allowing potential documentation or coding issues to be identified and addressed upstream. This reduces reliance on post-bill corrections and helps improve first-pass accuracy."
Nebraska Methodist is able to be paid sooner, with corrections being handled pre-bill.
RESULTS
On the back end, claim status automation results have been significant.
By automating high-volume, repetitive claim status activity, Nebraska Methodist has freed up the equivalent of 24 full-time employees over the last two years. Previously, staff spent a significant portion of their day manually checking payer portals for routine updates. With AI handling this work at scale, 71% of accounts were removed from staff queues, Francis reported.
"Rather than reducing headcount, we intentionally redeployed those teams to higher-value work, including complex claim resolution, denial management and payer-specific follow-up," he explained. "This improved productivity while lowering the overall cost to collect.
"Furthermore, we have added more than 100 beds without increasing FTEs in the business office," he added.
And there were potent results in mid-cycle coding accuracy and pre-bill impact.
"Separately, AI-driven coding support allowed us to identify issues before claims were submitted rather than reacting to denials months later," he reported. "We worked with Akasa to deploy its Coding Optimizer, designed to support human coders by reviewing increasingly complex medical records and surfacing potential opportunities for closer review.
"The AI does not replace coder judgment – it instead acts as a consistent second set of eyes, flagging areas that may warrant attention and providing supporting evidence," he continued. "All recommendations flow through a human-in-the-loop review process, reinforcing coder confidence while meeting strict compliance standards."
The organization already has realized more than $2 million in gross revenue, achieved a cumulative CMI lift of +1.17, and flagged more than 1,800 additional quality indicators, he added.
"The biggest impact has been confidence – confidence the care delivered is being accurately reflected, defended and reimbursed in an increasingly aggressive payer environment," he concluded.
Follow Bill's health IT coverage on LinkedIn: Bill Siwicki
Email him: bsiwicki@himss.org
Healthcare IT News is a HIMSS Media publication.
WATCH NOW: How can an IT exec become a chief AI officer – and then work with the C-suite
