Deutsche Telekom

Solution

Deutsche Telekom built and implemented MINDR as a multi-agent network diagnostics and remediation system with Google Cloud. The deployment extends the production RAN Guardian Agent, which already runs live in Germany, into an end-to-end service operations architecture. Specialized agents ingest network signals and event data from multiple network domains, then correlate them using Google Cloud's Autonomous Network Operations framework together with Google's Gemini models on Vertex AI. The agents create a real-time cross-domain view of service performance, support root-cause analysis, and trigger controlled remediation actions. In production, RAN Guardian monitors traffic surges and major events, pre-checks affected mobile sites, and autonomously adjusts network settings when loads spike. The source does not describe the persistent storage layer, but it clearly shows an AI serving layer for explainable diagnostics and operator-facing or automated remediation across Deutsche Telekom's network operations.

Data flow

Signals and operational data from the RAN, transport, and core network domains are collected by specialized agents, correlated end to end using the Autonomous Network Operations framework with Gemini models on Vertex AI, assembled into a real-time service-performance view for root-cause analysis, and used to drive explainable recommendations or autonomous remediation actions such as network adjustments at affected mobile sites.