The industry conversation around healthcare AI has shifted. Health system leaders are no longer being asked whether to adopt the technology; they’re being held accountable for what adoption actually delivers. That accountability often rests on the answer to a simple question: does the tool reduce friction or create it?
“We see so many great ideas that end up being dead on arrival because they’re not able to complement existing, strenuous workflows that our clinicians are already facing,” said Jason Szczuka, chief digital officer of Cincinnati-based Bon Secours Mercy Health and president of Accrete Health Partners, the health system’s digital venture arm, in a recent “Becker’s Healthcare Podcast” episode.
The standard has real consequences. Clinicians are stretched, and tools that promise efficiency while creating new burdens tend to get abandoned regardless of how they performed in a pilot. That tension fuels much of the gap between what health systems have invested in AI and the hard return.
“The most important problem is moving from AI as a set of promising tools to AI as a repeatable operating capacity,” said Mr. Szczuka. “That means identifying your highest friction workflows, making sure the technology is safe and usable, and then importantly holding teams accountable for measuring the improvements that it does or does not deliver. In healthcare, technology only matters if it helps someone make a better decision, get care more easily, or spend less time doing work that could be automated.”
That standard has grown more urgent as financial pressure narrows health systems’ tolerance for speculative investment. Health systems across the country are working to turn AI pilots into measurable outcomes as budget cycles tighten and earlier bets on AI come up for review. Organizations that accumulated proofs of concept without clear ownership or defined outcomes are beginning to feel the pressure.
“The winners are not going to be the organizations with the most pilots out and going,” he said. “They’re going to be the ones that connect AI and digital tools to real operating problems with clear ownership and measurable outcomes.”
Health systems are governing AI deployment more deliberately, moving away from pilot accumulation toward structured governance and explicit value thresholds for deployment, which organizations that built early AI programs without accountability frameworks have since had to reckon with.
Bon Secours Mercy Health’s entry point for any technology starts with the operating problem, not the product. The health system uses AI, digital ventures and operating partnerships to improve access, experiences and efficiencies, but only where those improvements can be embedded inside the system rather than layered on top of it.
“It can’t become just another dashboard layer,” Mr. Szczuka said. “It has to reduce friction. It has to improve decisions, and at the end of the day, it has to create capacity that can be reinvested and redeployed into the higher quality, lower cost delivery of healthcare.”
That philosophy runs through both sides of his role — the operating mandate as digital chief and the investment discipline he applies through Accrete Health Partners. Innovation that doesn’t connect to a measurable operational outcome fails to support the health system’s mission.
“We’re not pursuing technology just for its own sake,” Mr. Szczuka said. “We’re asking how can digital tools help people access our care better, help our clinicians and our operators work smarter, and most importantly, help our ministry deliver on its mission more effectively.”
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.