Kaiser Permanente built an early-warning model for unplanned ICU transfers on H2O.ai.

Unplanned ICU transfers carry 2-5x higher mortality and 8-12 longer hospital days, so Kaiser needed earlier warning. It used H2O.ai's machine learning platform to build the predictive system. The page quotes the team on faster processing and no longer needing to sample data; it gives no outcome figures.

H2O.ai is a machine learning and document AI vendor, used by health systems and pharma for prediction and automating faxed and referral documents

This sits in the Build stage of the AI use-case lifecycle: Named examples from the foundation-model labs and platforms pharma teams use to actually construct, train and generate the model or application.