For one day, startups, universities, and research institutes came together in Mannheim to talk about Industrial AI. Sounds like the beginning of a joke—or perhaps the beginning of a fellowship.
In this episode, Christian and Michael look back at the AI Research Day, which they co-hosted with SysDICE. The big question: What does it actually take to make AI work in complex industrial environments?
One thing quickly became clear: Industrial AI is all about context. Models need more than data—they need to understand the processes, systems, and technical relationships behind it. Model harnesses might become a crucial piece of the puzzle, while reference architectures could provide the structure needed to make all of this work together. Yes, we know: reference architectures can get a little academic. But maybe this time, it’s worth the risk.
The most concrete outcome of the day? The participants decided to write a joint whitepaper on Industrial AI. With one central idea at its heart: the Digital Thread as the key to giving AI the context it needs to understand deep technical and industrial processes.
So, is this the beginning of a new fellowship? Or just another anti-pattern waiting to happen?
Let’s find out.
Podcast: Play in new window | Download ()
