Deutsche Telekom broadens agentic AI across network operations
Deutsche Telekom is extending its agentic AI work beyond radio access network operations into broader, cross-domain network automation. Its RAN Guardian system is already active in Germany, using three agents to detect events, gauge impact and apply selected fixes. The operator is developing MINDR as a multi-agent platform that can correlate signals, diagnose problems and act within guardrails, with human approval retained for critical changes.
Deutsche Telekom is moving its agentic AI efforts from radio access network optimization toward automation that spans multiple network domains. RAN Guardian is already operating in Germany, where three agents identify events, evaluate their likely network impact, and carry out chosen corrective actions.
The operator is also developing MINDR, a multi-agent platform intended to connect signals across domains, diagnose incidents, and execute remedies inside defined limits. Critical changes would still require human sign-off, according to the company’s presentation at its AI Investor Day.
As carriers adopt agentic automation, mobile users could see fewer congestion-related disruptions during large gatherings, while network engineers may shift from manual troubleshooting to oversight and guardrail design. The approach may improve service reliability and speed, but its reliance on automated decisions means customers and operators could be affected by errors or opaque remediation unless human approval and monitoring remain robust.