How Escalation Rules Protect Your Caller Experience

Not every call can or should be handled entirely by AI. Escalation rules give you precise control over when a call gets transferred to a human, and how that handoff happens.

What Escalation Rules Are

An escalation rule has three parts:

1. Condition — what triggers the escalation
2. Action — what happens (transfer, callback, voicemail, ticket)
3. Agent behavior — what the agent says during the handoff

Available Conditions

• Caller request — the caller explicitly asks for a human ("Let me talk to a person")
• Sentiment threshold — the caller's frustration level exceeds a threshold (e.g., sentiment drops below 3)
• Keywords — specific words or phrases trigger escalation ("lawsuit", "complaint", "manager")
• Duration — the call exceeds a time limit without resolution
• Repeated failure — the agent fails to help after multiple attempts
• Topic match — the detected intent matches a configured topic (e.g., "legal" intent always escalates)
• Custom conditions — combine multiple conditions with AND/OR logic

Available Actions

• Warm transfer — the agent introduces the caller and context to the team member before connecting them. "I'm going to connect you with Sarah, who handles billing. I'll let her know you're asking about a charge on your last statement."
• Scheduled callback — the agent collects the caller's preferred callback time and creates a callback request for the team
• Voicemail — route to voicemail with the team member's greeting
• Support ticket — create a ticket with the full call context for follow-up

Agent Behavior During Escalation

You configure what the agent says when escalating:

• "Let me connect you with someone who can help with that directly."
• "I want to make sure you get the best help on this — let me transfer you to our billing team."
• "I understand this needs personal attention. Let me get you to the right person."

The agent also provides context to the receiving team member — a brief summary of what the caller needs, what's already been discussed, and why the call is being transferred.

Escalation Logging

Every escalation is logged with:

• Which rule triggered
• The condition that matched
• The action taken
• The transcript up to that point
• The outcome (successful transfer, voicemail, callback scheduled)

This lets you review escalation patterns — if the same rule triggers frequently, you might need to add more knowledge to the agent's base so it can handle those calls directly.

Zero Escalation Mode

For businesses that want the agent to handle everything without transfers, you can configure zero escalation mode. The agent will never transfer — instead, it always offers to take a message or schedule a callback. This is useful for after-hours operation or businesses without staff available for transfers.

Location-Specific Rules

Escalation rules can be scoped to specific locations. Your downtown office might escalate emergency calls to an on-site provider, while your satellite office creates a callback request for the same situation.

You can ask Allison to add, update, enable, disable, or delete escalation rules — through the dashboard chat or by calling Allison. Complex multi-condition rules are best configured through the dashboard.