• Controller Virtualization: What real opportunities – and risks – do you see in decoupling PLCs from physical racks?
• Cyber Security: How can IT-grade security principles be applied effectively to virtualized machine controllers?
• Access Control: In the past, plant managers opened the gate; today, only IT can open the firewall. How should responsibilities be rebalanced?
• Strategic Perspective: Is machine programming already recognized in your company as a new B2B digital interface, and is IT involved early enough?

With the rapid evolution of large language models (LLMs) and intelligent automation technologies, AI agents are becoming powerful tools in the mechanical engineering sector. These agents can make autonomous decisions, analyze complex data, and interact with machines and systems in real time—transforming traditional processes across maintenance, production, sales, and service. This panel discussion explores how AI agents are being integrated into Industry 4.0, what their current capabilities are, and what ethical, technical, and operational challenges still stand in the way of widespread adoption. Are we at the dawn of a new era where humans and AI co-pilot the factory of the future?

• What distinguishes AI agents from traditional automation systems and conventional AI applications like chatbots?
• Which use cases in mechanical engineering are showing the greatest value from deploying AI agents today?
• How can AI agents be safely integrated into critical manufacturing processes without compromising reliability and control?
• What are the main challenges in training and governing industrial AI agents—especially in terms of data quality and compliance (e.g., GDPR)?
• How will the role of engineers and operators evolve as AI agents take on more autonomous tasks? Will they replace, augment, or collaborate with human expertise?

Industrial digitalization is not held back by a lack of technology, but by fragmentation. Every layer, from IoT sensors and edge processors to controllers, applications, CAD, PLM, ERP, and MES produces its own data models and semantics. Across suppliers and partners throughout the supply chain, the standards multiply. These isolated silos create hidden barriers, making integration brittle, expensive, and slow.

SEW-EURODRIVE addresses this challenge with a Data Integration Framework – a structured methodology that thrives beyond toolchains or platforms. It establishes a true data plane that unifies semantics across IT and OT, powered by Knowledge Graphs and Industry 4.0 Asset Administration Shells. This approach enables Digital Twins as living representations of types and assets, ensuring interoperability, governance, and lifecycle integration.

By reframing the problem as a systemic barrier, this presentation shows how mastering semantic complexity is key to unlocking development speed, regulatory compliance and new industrial business models strengthening Europe’s competitiveness.

• Industrial data is fragmented across IoT sensors, edge processors, controllers, apps, and IT systems (CAD, PLM, ERP, MES).
• This fragmentation creates systemic barriers that block fast, reliable, and scalable integration.
• Without solving this, Digital Twins remain pilots, and cross-domain innovation, compliance, and lifecycle traceability fail.
• SEW-EURODRIVE applies a Data Integration Framework that builds a unifying data plane across IT and OT.
• Core enablers: Knowledge Graphs for semantic linking and Asset Administration Shells (AAS) for Industry 4.0 compliance.
• Silos are not destroyed but integrated through explicit, machine-readable domain knowledge.
• Governance ensures trust, scalability, and resilience.
• Strategic advantages: semantic & technical interoperability, operational efficiency, compliance, and cyber resilience.