Our Take
Real operational gains from boring AI automation, not breakthrough technology but competent execution of workflow improvements.
Why it matters
Revenue cycle management remains one of healthcare's most concrete AI success stories, with measurable error reduction and faster cash flow that directly impacts hospital finances.
Do this week
RCM leaders: audit your current duplicate work patterns this week so you can identify which manual processes AI can automate first.
Northwell Health cuts RCM errors 40% over five years
Northwell Health Labs implemented AI in its revenue cycle management system over a five-year period, achieving a 40% improvement in error rates (company-reported). Joe Accurso, vice president of Revenue Cycle at Northwell Health Labs, deployed the system in what he calls "digestible chunks" rather than attempting a wholesale replacement.
The AI system created automated workflows and work queues that eliminated duplicate work on insurance claims. Staff previously found multiple people working the same claim, creating inefficiency and frustration. The new system routes claims automatically and prevents overlap.
"AI was able to find and address errors that couldn't be found through the manual process," according to the implementation. Cash collection accelerated as a result of the error reduction and workflow automation.
Management overhead shifted from reports to productivity
The change forced management adaptation. Previously, managers created manual reports to distribute work. AI work queues eliminated this step, removing what Accurso calls staff excuses: "I didn't get my report yet." Managers had to learn to manage higher productivity standards directly.
"We were able to create more efficiencies which removed the excuses," Accurso said. "I think the staff probably appreciated that more than the management team."
The staff adjustment period involved resistance to change, but satisfaction increased once the duplicate work disappeared. The system has been running for approximately five years, indicating operational stability.
Start small, vet vendors thoroughly
Accurso recommends implementing AI solutions incrementally rather than building massive systems that delay results. His approach focuses on getting benefits quickly from smaller deployments before expanding scope.
Vendor evaluation requires thorough vetting of processes, relationships, and track records with health systems specifically. The revenue cycle management application appears to target workflow automation rather than clinical decision-making, reducing regulatory complexity.
Healthcare finance teams should expect management process changes alongside the technology deployment. The elimination of manual reporting creates accountability that some management layers may resist.