Ask what's wrong with your systems. The agent reads your cluster, metrics and alerts, then explains the root cause.
Describe the problem in plain words. The agent checks pods, events, logs and metrics, and shows every step it takes.
New alerts are investigated automatically.