Nearly half of all paid work hours are tied to physical tasks, from moving and making things to inspecting equipment and transporting goods. Today, much of that work remains beyond the reach of traditional automation. Physical AI could begin to change that, expanding the range of tasks machines can perform reliably, safely, and economically. That means more than robots that can outpace Usain Bolt; it means intelligent, adaptable systems that “sense, decide, and act based on the physical world around them,” according to McKinsey Senior Partner Mark Patel. And that physical world is “full of corner cases.”

On this episode of The McKinsey Podcast, Patel and McKinsey Partner Ani Kelkar join Global Editorial Director Lucia Rahilly to explain what physical AI looks like, where it could create value, and what leaders should know now to begin taking advantage of its potential for impact.

The McKinsey Podcast is cohosted by Lucia Rahilly and Roberta Fusaro.

This episode has been adapted from our McKinsey Live webinar series.