Utilization management has long lived inside the EMR by default, not by design. For many health systems, that means physician advisors and UR nurses work from different views of the same patient, reviews go unprioritized, and cases that should be appealed to inpatient status simply aren’t caught in time. The University of Tennessee Medical Center (UTMC) found itself in exactly that position in 2024 — and its response offers a window into how one system is rethinking UM from the ground up.
UTMC’s observation rate was running high, and too few peer-to-peer reviews were getting completed to successfully appeal cases that warranted inpatient status. Leadership traced both problems back to the same root cause: a UM process built entirely on top of an EMR never designed for it.
Mark Reddington, DO, MSc, CHCQM-PHYADV, joined UTMC as Medical Director of Utilization Management shortly before the health system began implementing a new UM vendor, recognizing the downstream impact of inefficiencies in UM. Leadership, he said, wanted a company with staying power in the mid-revenue cycle space, a demonstrated ROI track record, and the ability to support peer-to-peer reviews directly. UTMC landed on Xsolis, an AI-native healthcare technology company whose Dragonfly platform applies predictive analytics and generative AI across the utilization review process, rather than a services vendor with UM software or AI tools layered on top.
A rollout in two phases
UTMC and Xsolis didn’t implement everything at once. The first phase, which went live in March 2025, gave UTMC’s physician advisors access to Xsolis’ Physician Advisor Services along with the Dragonfly Advise platform — a tool Dr. Reddington calls “Dragonfly Light” — which uses AI and machine learning to help physician advisors manage case volumes and prioritize escalated cases.
That first phase produced enough of a return that UTMC expanded the partnership. Phase two, launched in November 2025, brought the health system onto the full Dragonfly platform, adding Dragonfly Utilize, an AI-driven UM tool, and GenAI Initial Review, a generative AI feature that drafts clinical summaries to speed up the review process.
Getting UR nurses and physician advisors on the same page
More than six months into the full rollout, Dr. Reddington points to one change above the rest: physician advisors and UR nurses are now working from the same platform, in real time, instead of reconstructing a shared picture from separate EMR views. That shared view is powered by Xsolis’ Care Level Score®, predictive analytics intended to inform clinical judgment when determining medical necessity rather than replace it with rigid criteria.
“EMRs aren’t built for utilization management, which creates significant data gaps,” Dr. Reddington said. He described Dragonfly less as an add-on tool than as the EMR’s missing counterpart for UM specifically: a system built to do for utilization review what the EMR does for clinical documentation. That distinction shows up in daily use, he said, in the form of clear direction on which patients need review most urgently, whether the concern is a high observation rate, missed inpatient-only procedures, or short-stay Medicare Advantage cases — prioritization he said an EMR alone was never built to surface.
Dr. Reddington said UR nurses are moving through cases faster and missing fewer of them since the full platform went live, across inpatient-only cases as well as MA plans and short stays. He credits the shift as much to growing staff trust in the AI tools as to the tools themselves.
An account team that keeps working after go-live
Dr. Reddington singled out Xsolis’ account team as a distinguishing factor in the partnership, separate from the technology itself. “Our account team meets us where we are, partnering with us and using reporting to identify opportunities for improvement,” he said, adding that the relationship has felt consistent “from the initial sales demos” onward.
UTMC and Xsolis are now using that reporting relationship to plan next steps, translating performance data into action plans with the account team. With GenAI Initial Review now live, additional time savings and smoother interactions with payers for both the UR and physician advisor teams can be expected.
To hear more from Dr. Reddington on UTMC’s experience, watch the full video series here.
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.
