Beyond Simulation: How AI Supercharges the Digital Twin

This video features Dominil Zettler, VP Horizontal Management Simulation for Industry at Siemens, explaining how digital twins and industrial AI work together — recorded on the sidelines of Hannover Messe 2026. Zettler opens by defining the digital twin as a virtual representation of a physical asset, process, or system, noting the concept has been established in industry for around a decade. He distinguishes three core dimensions: the digital twin of product, the digital twin of production, and the digital twin of performance — explaining that combining all three forms a comprehensive digital twin, with different levels of granularity within each area. The presentation then draws a clear distinction between consumer AI and industrial AI, outlining what industrial environments specifically require: safety, reliability, consistency, and the ability to manage complex dependencies repeatedly at the same quality level. Zettler describes the AI architecture relevant to industry, covering industrial foundational models, agent orchestration, large language model-based copilots, and physical AI — the latter directed at more autonomous production through training robots and AGVs using sensor and video input. The central example is a conveyor sorting system. Zettler explains how an AI algorithm is trained entirely within a simulated, virtual environment — running up to 20 million cycles — before being deployed to a real-world conveyor. The outcome is a self-adapting system that aligns and sorts parcels across multiple belt configurations without hard-coded instructions. Zettler closes by noting that the pairing of digital twins and AI extends across different industries and applications, with both technologies reinforcing each other through testing, training, and validation. You can explore the full potentials of the Digital Twin by visiting this page: https://sie.ag/69nY5u