Teach the AI how your team triages alerts
A policy is a plain-English rule you write — no code. Every enabled policy is handed to the AI on every alert analysis as strong guidance. The AI follows it where it applies and cites it in its reasoning, or explains in the audit trail when it overrides one.
-
1
Write a rule in plain English
e.g. “Dismiss alerts under $500 from licensed VN exchanges.”
-
2
Switch it on
Enabled policies go to the AI on every analysis. Toggle off to pause one without deleting it.
-
3
The AI applies it
It follows matching policies as strong guidance and names them in its reasoning — every change is versioned.
No policies yet
Alerts are analysed with the built-in logic only. New here? Start from the rules our team triages with today — tap one to load it into the editor, then tweak and save: