Support & Operations
Escalation Rules
Conditions that trigger an automatic handoff from AI to a human agent.
Escalation rules are the logic that determines when an AI chat agent should stop handling a conversation and transfer it to a human. Common triggers include: the visitor explicitly asking for a human, negative sentiment detected in the conversation, the agent failing to answer the same question twice, or specific keywords like 'cancel', 'refund', or 'legal'. Well-defined escalation rules are the difference between a helpful AI agent and one that frustrates users by refusing to escalate appropriately.