Escalations
The Escalations module gives you an AI escalation manager: an employee whose job is to watch for problem patterns across the team and bring them to a human's attention before they pile up unseen. It exists for the situation where issues quietly stack up over a busy day and nobody flags them until you happen to look.
The escalation manager
Pick any AI employee as the escalation manager. The reviewer is a natural choice, since they already see everyone's work. They watch for patterns that match your escalation rules and, when one fires, notify you in the notification center, in the Updates feed, and on Slack. Because they are an AI employee and not a rigid automation, they reason about whether a situation really matches a rule.
Standard escalation rules
A set of rules ships ready to enable:
- The same blocker across three tasks, where an employee keeps hitting the same wall.
- A task pending for 24 hours or more with no human response.
- An agent failing to deliver work two or more times.
- A blocker affecting a deal or revenue pipeline, treated as critical.
Each escalation carries a severity of low, medium, high, or critical.
Custom rules
Add your own rules with a name, description, and severity. Because the manager is an AI employee, you can describe the rule in plain language and let them interpret it, rather than writing rigid conditions.
Notification routing and method
Decide who gets notified, whether the organization owner, admins, or specific people, and choose the contact method: Slack DM, the updates feed, or both.