Enterprise Apps Unpacked
View All Series Articles
Podcast

Inside SAP R&D on the convergence of agentic and physical AI

In this podcast, Yaad Oren, SAP's head of research and innovation and managing director of SAP Labs U.S., says smarter, autonomous physical AI devices are poised for business use.

David Essex
By
Published: 06 Apr 2026

Agentic and physical AI have similar capabilities but apply them in vastly different worlds. Both are artificial intelligences that can learn and have some autonomy to make decisions and act in their environments. But for agentic AI, the environment is usually digital, while physical AI is embodied in devices like drones, factory robots and self-driving cars that can respond to and act on the physical world.

Some experts say physical AI is reaching a tipping point as agentic AI brings more agency to robots than ever before, and the lines between real and virtual worlds break down.

This convergence of agentic and physical AI is a focus of R&D at SAP as the ERP vendor investigates ways smart robots and other physical AI devices can work with business applications to make entire enterprises more intelligent and automated.

In this episode of Enterprise Apps Unpacked, Yaad Oren describes SAP's work in this area and argues that software is the reason physical AI will move from pilots into core business operations in 2026. Oren is SAP's global head of research and innovation and managing director of SAP Labs U.S., a dual role that enables him to explore emerging technologies, build early-stage products and ultimately bring the innovations to market.

Supervising robots with bigger brains

Oren explained how the current state of the art differs from the initial years of generative AI, with its emphasis on large language models and "understanding" human language well enough to analyze it and generate realistic, reasonably accurate content.

"When we are applying current LLM technology to physical AI, it's about the robot having a brain, if you will, that enables it to learn tasks and understand context," Oren said. Now when a field technician asks a robot to pick and place an item, it knows how to interpret the task and carry it out. "Just like you talk with a prompt and get answers today from an LLM, you can talk to or prompt a robot."

Yaad Oren, SAP global had of research and innovationYaad Oren

Today's LLMs are also helping warehouse robots perceive their environments more accurately and move through them efficiently, according to Oren. For example, Gen AI is making robot vision sharper and improving location awareness through remote sensing technologies like LiDAR (light detection and ranging), which uses lasers to measure distance and detect objects.

Recent Gen AI developments have essentially worked to give robots bigger brains, and now agentic AI enables them to act more autonomously, according to Oren. Multi-agent orchestration, a hot topic as vendors including SAP and Oracle work to build teams of agents capable of taking over entire workflows in ERP, has its analog in physical AI. He gave the example of a person who needs a screwdriver. They can ask another person who is closer to one to bring it to them. Recent AI has advanced "miles ahead" to give robots comparable location awareness and decision-making abilities, he said.

Adding AI to geofencing, a class of wireless technologies for establishing virtual boundaries, is another emerging application of physical AI. "Now that we are talking about a physical entity, there is a lot of work we need to do for safety and security," Oren said. "You don't want a robotic arm to slap an employee's face."

Other topics discussed in the podcast include the following:

  • Why enterprise warehouse management is a promising early application.
  • The role of enterprise software, especially ERP, in physical AI.
  • Risks and challenges business and IT leaders need to know.
  • How physical AI changes related older technologies such as IoT.

David Essex is an industry editor who creates in-depth content on enterprise applications, emerging technology and market trends for several Informa TechTarget websites.

Dig Deeper on ERP administration and management