Sharing this because I think it's a useful case study for anyone thinking about applying multi-agent AI to network ops and infrastructure automation.

The problem:

A Florida telecom operator was managing 254 Juniper switches manually. Every provisioning job required engineers to SSH in, run diagnostics, identify issues, generate fix commands, apply them, and update the Jira ticket. End-to-end: 1,016 engineer-hours across their fleet.

What we built:

A 3-agent system on our platform:

Master Orchestrator — receives network alerts (email / syslog), parses and classifies the issue, coordinates child agents, manages Jira integration

Syslog Parser Agent — extracts device identity, error codes, severity from raw syslog messages

Command Generator Agent — generates precise debug and remediation commands based on device context and issue type

Human-in-the-loop checkpoint: engineer reviews and approves the fix before execution.…