Health systems are not running out of AI products to consider. The harder problem is deciding which ones warrant the attention and in what order.
The market for AI tools has outpaced the governance structures most organizations built to assess them. Established vendors are rebranding legacy features as AI-enabled while purpose-built startups compete across clinical, operational and administrative use cases. Digital health leaders across the country are reckoning with how to build the right structure to absorb that volume without creating a bottleneck that slows legitimate innovation.
Cheng-Kai Kao, MD, chief medical information officer at UChicago Medicine in Chicago, said the pressure is constant.
“There’s AI products everywhere,” Dr. Kao said. “And even some of the older products, they call themselves smart or call themselves AI enabled, and there’s just a ton of AI products we’re dealing with or evaluating every day.”
UChicago Medicine’s answer was to build a single entry point for every AI submission from external vendors and internal teams alike. The process runs every product through cybersecurity review, finance review, and use-case assessment under a unified workflow. A dedicated team reviews submissions weekly and routes each one to the appropriate stakeholders based on what the tool does and who it affects.
The intake process also makes an upfront determination that many organizations defer until late in an evaluation: whether a pilot is needed at all, and if so, what it should look like. By resolving that question at the intake stage, the process removes a common source of ambiguity and delay.
The mechanism that drives those decisions is risk classification. Dr. Kao said the level of validation required before a tool goes live is calibrated to the nature of the application.
“Something that’s more administrative and there’s always a human in the loop, probably lower risk, you can probably turn it on and just have close feedback with the frontline staff to understand whether it works or not,” Dr. Kao said.
The calculus changes for tools that reach patients more directly. When an application operates without a human reviewer in the decision path, and when a patient is on the receiving end of that output, the validation threshold is substantially higher. Those tools move through a more rigorous assessment before any deployment, a deliberate trade of speed for safety.
Building that classification into the intake process from the start avoids a common failure mode: discovering a tool’s risk profile partway through an evaluation and having to restart the governance process. The level of scrutiny a submission will face is determined at the entry point, not negotiated after evaluation is already underway.
The underlying principle is that governance intensity should match actual risk. An administrative tool that surfaces information for a staff member is structurally different from an application that delivers a recommendation directly to a patient, treating them identically wastes evaluation capacity on one end while understating risk on the other.
Becker’s 3rd Annual Fall Chief Pharmacy Officer Summit is the top tier event that the people leading the future of pharmacy convene. This 3-day event is built specifically for Chief Pharmacy Officers and senior pharmacy leaders. The sessions examine the workforce, financial and clinical priorities shaping pharmacy today. Becker’s is where candid discussions flourish and new ideas thrive. Becker's 3rd Annual Fall Chief Pharmacy Officer Summit
