Jeremias Dötterl showcases explainable agentic AI troubleshooting for optical networks at ECOC 2026
Demo will show how integrating data references into agentic AI improves root-cause analysis and operator confidence
- Speakers: Jeremias Dötterl, data scientist in advanced technology at Adtran
- Topics: Preventing data hallucinations in agentic AI for optical network troubleshooting
- When: Tuesday, September 22, 11 a.m. CEST
- Where: ECOC 2026, The Malaga Trade Fair and Conference Center (FYCMA), Málaga, Spain https://ecoc2026.org/ecoc2026
As AI takes on a greater role in network operations, operators need confidence that automated systems can deliver accurate, verifiable results. For optical network troubleshooting, preventing AI hallucinations is critical to reliable root-cause identification, operational efficiency and trust in automated decision-making.
In this demonstration, Jeremias Dötterl will showcase an agentic AI system for automated troubleshooting in optical networks. He will show how integrating referenced network data directly into the agent architecture helps prevent hallucinations during fault analysis and supports more transparent, verifiable reasoning. Attendees will learn how grounding AI decisions in network data can improve diagnostic accuracy and strengthen confidence in AI-assisted operations.