Artificial intelligence-driven predictive tool for digital platforms

Abstract

Examples described herein relate to an artificial intelligence-driven tool for digital physical therapy. First data indicative of a baseline condition of a first user is accessed. Second data is collected for each of a plurality of sessions of a digital physical therapy program. A first device associated with the first user tracks motion of the first user during each session. Input data based on the first data and the second data is provided to a machine learning classifier to cause generation of output data indicative of a predicted outcome of the digital physical therapy program. The predicted outcome is processed to detect that the first user is predicted not to meet a predetermined improvement threshold. An alert is generated and presented at a second device associated with a second user administering the digital physical therapy program.