Physical AI in practice: Technical foundations that fuel human-machine interactions

In this post, we explore the complete development lifecycle of physical AI—from data collection and model training to edge deployment—and examine how these intelligent systems learn to understand, reason, and interact with the physical world through continuous feedback loops. We illustrate this workflow through Diligent Robotics’ Moxi, a mobile manipulation robot that has completed over 1.2 million deliveries in hospitals, saving nearly 600,000 hours for clinical staff while transforming healthcare logistics and returning valuable time to patient care.