HonorHealth EVP Dr. Jim Whitfill explains how his system weighs returns on AI spending and why he favors guardrails on token costs. Watch below or on YouTube.
HonorHealth can require its most expensive AI resources to project a 4-to-1 financial return before work starts, and it holds other AI tools to lighter tests. Dr. Jim Whitfill, EVP, Strategy, Transformation and Digital at the nine-hospital Arizona system, laid out the approach on a recent episode of the healthsystemCIO Show. AI is moving so fast that it is hard to know exactly where its value will come from, he said, so HonorHealth spreads its AI investments across four lanes. Lanes that produce results get more money. Lanes that stall, meanwhile, are paused, rethought and restarted.
The approach borrows from the way venture capitalists build a portfolio, Whitfill said. In his rough illustration, a firm backing 10 companies expects one to hit big, two to do fine and seven to fail, so it spreads the risk. HonorHealth’s first lane, for example, gives operators Microsoft Copilot to find efficiencies in their daily work. The second draws on AI that Epic and Workday, its EHR and ERP vendors, build into their products. In the third, HonorHealth builds its own AI on Google’s cloud, where its analytics and data already sit.
The fourth lane is a partnership with Qventus that embeds five AI engineers at HonorHealth. They take on complex work with potentially high returns, such as referral management, which runs across several systems. Qventus also supplies the commercial software HonorHealth uses for patient throughput and operating room utilization.
Source link







