16th FVI Expert Breakfast

Topic: "AI Directly from the PLC" – WAGO shows how easily machines can learn.

In this session, we left the theoretical level and went directly into the control cabinet with WAGO (Jürgen Pfeifer, Michael Trost). The central question: Does AI always have to be cloud-based, expensive, and complicated? Answer: No.

• The "No-Code" Revolution in the Control Cabinet: Michael Trost demonstrated live how to easily activate AI modules (e.g., anomaly detection) on a WAGO controller via drag-and-drop. No programming, no data scientist needed. This dramatically lowers the entry barrier.
• The "Clint Eastwood" Principle (Shoot first, ask questions later): Marcel Hahn and Christian Kesper (Bayer) advocated for collecting data first (in a time-series database), even if you don't yet know what for. When the error occurs, it's too late to generate data. Storage space costs nothing anymore, missing data costs downtime.
• Harvesting Low Hanging Fruits: You don't have to build the "Digital Twin" for the entire factory right away. A simple dashboard (Grafana) that shows "Filter clogged" already saves 50% of troubleshooting time. It's about pragmatic help, not science fiction.
• The SAP Trap: Karl (IFM) reported on customers who buy expensive sensors only to send a dumb counter reading to SAP. That's wasteful. Real intelligence arises when data is analyzed on-site (Edge Computing) and directly provides the worker with action recommendations ("Grease the bearing!").
• Courage to Take Risks: Christian (Bayer) called for abandoning the "German error culture." We discuss innovations to death before trying them. Just do it, risk mistakes, learn from them. The WAGO system allows exactly this playful experimentation without risk to the plant.

Classification: Edge AI as the Foundation for ADAM

This episode is the perfect technical complement to our software strategy.

• ADAM needs data: ADAM is the "head," but the PLC (WAGO) is the "nervous system." When WAGO detects the anomaly ("Vibration high"), ADAM can provide the worker with the solution ("Here is the guide for bearing replacement"). The combination of Edge AI (WAGO) and GenAI (ADAM) is unbeatable.
• Democratization of AI: WAGO makes AI usable for the PLC programmer. ADAM makes AI usable for the maintenance technician (via voice). Both pursue the same goal: hide complexity, maximize benefit.
• Sovereignty: The demonstrated solution runs locally (on the Edge Controller), not in the cloud. This perfectly supports our argument for data sovereignty. The customer retains control.

Conclusion: AI is no longer magic. It is now a component in the control cabinet, as simple as a fuse. Anyone not using it now is acting negligently.