Disconnected and air-gapped mission environments often depend on fragmented command systems, sensor feeds, network telemetry, and security tools, increasing operator burden and slowing response. This session presents an agentic AI architecture for mission network observability that uses Splunk as the trusted local evidence layer, enriched by Cisco and infrastructure telemetry, to unify disparate data into an actionable operational picture.
A governed agent layer leverages RAG, LLMs, and controlled tool access to support contextual reasoning, guided investigation, and operator-ready summaries without requiring cloud connectivity or uniform network architectures. Designed around modular, auditable, human-in-the-loop principles, the approach supports CMOSS/SOSA-aligned integration while reducing cognitive load and improving situational awareness across heterogeneous mission environments.