Engineering Agents That Actually Engineer

Agentic Engineering is taking off. But when AI agents meet complex engineered systems, the question is how we move from reasoning to execution and verified outcomes. This requires an executable reasoning surface, controlled orchestration of agents and tools, and enterprise safeguards for functional safety, resilience, traceability and human accountability.

This session presents an Agentic Engineering platform approach emerging from digital.auto's Forge project, a collaboration among Bosch, n8n, NVIDIA, HPE, AVL, Kaizenics and RemotiveLabs. n8n provides the agentic backbone, connecting reasoning, execution, observation, testing and refinement. The architecture integrates enterprise workflows with embedded development, simulation and system-level diagnostics, while supporting fault recovery and human-in-the-loop gates.

Using the Forge as an automotive reference implementation, the talk shares practical patterns for connecting agents with executable vehicle models and engineering infrastructure. It offers a concrete path from agent demonstrations to trustworthy, continuous engineering of complex physical systems - in automotive and beyond.

• Learn a concrete architectural pattern for connecting AI agents to complex engineered systems, using n8n as the agentic backbone
• See how reasoning, execution, testing, and refinement are orchestrated across real embedded, simulation, and diagnostic tools
• Understand practical safeguards for making agentic systems trustworthy in production, including fault recovery, human-in-the-loop gates, and traceability
• Discover a transferable path from agent demos to reliable, continuous engineering of physical systems, using the automotive Forge project as a reference