How can AI agents effectively support industrial processes in the future — not just as assistant systems, but as a link connecting entire workflows from planning and engineering all the way to the shop floor? This question was the focus of the event “Smart Agents for Smart Factories,” which took place on June 9 at the CENTRUM INDUSTRIAL IT (CIIT).
The event was part of the KI-Akademie OWL and was organized by Prof. Dr. Markus Lange-Hegermann from inIT and Dr. Gesa Benndorf from Fraunhofer IOSB-INA. Around 120 participants from academia, industry, and the application sector came together to discuss opportunities, challenges, and concrete applications of AI agents in industrial contexts.
AI Agents: from research to industrial applications
In their keynote speeches, Dr. Chantal Sinnwell of Siemens Digital Industries Software and Anita Allen of n8n presented two key perspectives on the topic.
While Dr. Chantal Sinnwell highlighted the potential of AI agents in connected production environments, Anita Allen focused on the practical orchestration of agent-based workflows. The discussion centered on how different systems, applications, and data sources can be integrated in such a way that individual AI applications give rise to end-to-end, actionable processes.
Discussion on potential applications, safety, and acceptance
In addition to the keynote speeches, a poster session, a World Café, a LabTour through SmartFactoryOWL, and a closing panel discussion provided a variety of opportunities for professional exchange. Topics discussed included the tasks for which AI agents are already suitable today, how small and medium-sized enterprises in particular can successfully get started, and the role that security, governance, and acceptance play in the implementation process.
It became clear that AI agents can create added value above all where tasks are clearly structured, processes are easily definable, and data is readily available. At the same time, their deployment requires more than just technological capability. Crucial factors include suitable framework conditions, transparent decision-making, and integration that fits into existing industrial workflows.
inIT was represented with several poster presentations
Researchers from inIT also presented their latest work during the poster session. Krithiga Ramesh, Simon Jonas Leister, Julia Becker, Mortimer Dockhorn, and Steffen Fricke provided insights into research topics including industrial agent systems, security risk assessment, LLM-based assistance systems, and networked production processes.
Exchange as a driver of transfer
“AI agents open up new possibilities for making industrial processes smarter and more flexible. However, it is crucial that we do not view these technologies in isolation, but rather integrate them meaningfully into existing systems, data structures, and workflows. That is why it was particularly rewarding to bring together so many people from research, industry, and application at the Innovation Campus Lemgo. The exchange demonstrated how valuable direct dialogue is for developing viable solutions for industrial practice,” summarizes Prof. Dr. Markus Lange-Hegermann.