- Operate incidents (alerts → cases → evidence → response)
- Make incidents possible (connect sources/integrations so alerts flow)
The mental model (visual)
CoPilot becomes intuitive when you see it as two loops that share the same data:Admin / Platform loop
You make detection possible.
How alerts become cases (the shared pipeline)
1) Data arrives
Endpoints (Wazuh), API integrations (O365/Mimecast/Huntress/CrowdStrike), and syslog devices (FortiGate/PAN‑OS/ASA).
“Where do I click?” (quick map)
Operators
Incident Management → Alerts → Cases
Admins / engineers
Provisioning → Integrations → Connectors → Indices
Developers
Architecture → Data flows → Connectors
SOC operator / analyst
You spend most of your time in Incident Management. You care about:- Is this alert real?
- What’s the blast radius?
- What evidence do I need?
- What do I do next?
Admin / engineer
You care about data onboarding and reliability. You care about:- Are sources connected and healthy?
- Are parsing / streams / pipelines configured correctly?
- Are indices and retention behaving?
- Is each customer/tenant provisioned correctly?
Developer / AI agent
You are extending or modifying CoPilot. You care about:- Architecture + data flows
- Schema changes (Alembic as source of truth)
- Adding a connector safely
