Make Machines Smarter with Physical AI
Michael MacKenzie, Microsoft’s Vice President of Product Management for Azure IoT and Digital Operations, walks through how connected machines can move beyond monitoring to AI-guided action. He explains how Azure services support device connectivity, operational inference, and governance across IT and operational technology (OT).
From monitoring to learning machines
IoT gives businesses visibility into what happens in the physical world, but dashboards alone do not make machines smarter. MacKenzie describes three complementary roles:
- IoT tells you what happened, providing data from connected devices.
- Industrial AI tells you what should happen next, connecting that data to decisions.
- Physical AI makes it happen, bringing those decisions into machine execution.
The distinction from traditional automation is the learning loop: physical AI gets smarter with each execution, supporting the complex tasks machines already perform.
How the Azure services fit together
The video distinguishes the responsibilities of three Azure services:
| Service | Role |
|---|---|
| Azure IoT Hub | Connects and updates devices at global scale. |
| Azure IoT Operations | Curates data and runs inference in OT environments. |
| Azure Device Registry | Lets IT teams apply policy, role-based access control, and certificate management to OT devices. |
Together, these roles connect device management and operational data processing with the controls needed to bridge IT and OT.
Video chapters
- 0:00 — What IoT gave businesses in the physical world
- 0:40 — Why dashboards alone do not make machines smarter
- 1:05 — How IoT, industrial AI, and physical AI work together
- 1:35 — When to use Azure IoT Hub and Azure IoT Operations
- 2:20 — How Azure Device Registry bridges IT and OT
- 3:05 — How physical AI learns with every execution