Siemens isn’t rushing into the agentic AI wave, even though the buzz around it is growing fast.

The company sees potential, but also a lot of gaps that still need to be addressed before it can be used widely and reliably.

Speaking at Siemens Innovation Day 2026, Chief Technology and Strategy Officer Peter Koerte shared that the company is already experimenting with agentic AI internally. But he made it clear—they’re taking a careful approach. In his view, the technology is still at an early stage and needs better control, especially around security and decision-making.

Inside Siemens, teams are testing these ideas through something they call an “AI factory.” It’s essentially a space where employees can build applications using large language models and create their own AI agents for specific tasks. The setup even allows for building systems that can generate agents.

But despite all this progress, these systems aren’t fully independent yet. Human oversight is still a big part of the process. From checking outputs to ensuring safety, people are still very much in the loop.

At the same time, Siemens is steadily weaving AI into its existing products. Around 130 of its tools and solutions already use AI in some way, and that number is expected to grow. Rather than treating AI as a separate business, the company is embedding it across everything it builds.

To support this, Siemens has put together a sizable AI team—around 1,500 people, including researchers and engineers. It’s also expanding globally, with a new AI hub in Seattle that’s expected to grow over time.

When it comes to technology and infrastructure, Siemens isn’t trying to do everything alone. It’s working with major partners and taking a more cautious route with startups—focusing on investments instead of outright acquisitions.

One example is its investment in PhysicsX, which works on AI-driven simulations. But even here, the company admits it’s too early to know who will come out ahead in such a fast-moving space.

Siemens is also continuing its long-standing focus on innovation, filing thousands of patents every year, with a growing share linked to AI and data science.

Overall, the approach is quite balanced. Siemens is exploring agentic AI, building capabilities, and learning along the way—but it’s not in a hurry to scale something that still feels unfinished.

Also Read: Google Unveils Universal Commerce Protocol, New AI Tools to Strengthen Agentic Shopping