For a critical access hospital in rural Iowa, a failed technology bet isn’t an abstract risk. It’s a resource problem. Mahaska Health CIO Bob Berbeco said that reality shapes every decision that comes across his desk, especially when they’re AI driven.
The challenge isn’t finding AI solutions. It’s deciding which ones are worth pursuing when demand arrives from every direction and margins leave no room for error. Health system CIOs have made discipline and governance a defining priority as they work to scale AI beyond the pilot stage. Mr. Berbeco’s answer is a demand management framework that filters every incoming idea through a structured prioritization process tied directly to the organization’s strategic plan.
“Demand management is the method of IT,” Mr. Berbeco said during a recent episode of the “Becker’s Healthcare Podcast.” “Demand is everywhere. You’re literally in the middle and everything is coming at you. So you make that into a funnel. And as those ideas are coming your way, the potential solutions or even strategies, you bring those into the funnel, you weigh those, you look at the strategic roadmap for the organization, determine how those are prioritized there, and then you take them down the path of incorporating them or getting them to a solution state.”
The funnel doesn’t end at prioritization. Every idea that enters it must answer a harder question: how will the organization measure the return, and in what currency?
“It can be green dollars, which are true dollars, like actual money, or it could be blue dollars, which a lot of what you do with technology and healthcare tends to be blue dollar related,” Mr. Berbeco said. “So those could be reducing the amount of time in a workflow, reducing time in interactions to an EMR, those types of things.”
In practice, most healthcare technology investments — particularly in AI — generate blue dollar returns before they generate green ones — a gap health systems are actively working to close. Forcing every proposal to identify which type of return it’s targeting, and to define how that return will be measured, disciplines the evaluation before a dollar is spent and holds outcomes accountable after the fact.
Applied to Mahaska’s current AI roadmap, the framework has produced a defined portfolio: ambient documentation and visit summaries, predictive alerts, ED radiology triage pilots, telehealth no-show predictions and denial forecasting, with predictive staffing planned ahead. For ambient AI in particular, validating ROI before scaling enterprisewide has become a central challenge.
“We were doing experiments, and now we’re being more intentional, and we’re looking to expand on some of what we were experimenting with,” Mr. Berbeco said.
Underlying the framework is a clear test for where AI belongs in the operation. Mr. Berbeco calls it workflow-native AI — technology embedded directly in the processes where work happens, not layered on as a separate tool — and its success is measured by whether it reduces friction for the people using it.
“The biggest problem that we’re solving with AI right now is how do we make technology more useful and less burdensome for the people who use it, especially in rural health,” he said.
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.
