The steps
1
Connect a model
Pick the LLM the AI runs on - the Ankra default works out of the box, or bring your own key. Complete once an active provider is configured. See AI Provider and Use your own OpenRouter API key.
2
Autonomy and guardrails
Decide how far the AI may go on its own when an alert fires. Complete once the auto-remediation policy is enabled. See AI Autonomy.
3
Connections
Bind Slack or a delivery webhook so approvals and findings reach your team. Complete once Slack is connected or at least one delivery webhook is enabled. See AI Connections.
4
Proactive insights
Let the AI scan clusters continuously and raise findings before they page you - a single toggle, right in the checklist. See AI Insights.
5
Review and launch
A recap of the four steps, then straight to AI Home or the Inbox.