Default escalation triggers: positive reply, scheduling request, complaint, opt-out, unusual signal (legal mention, pricing pushback, request for an exec). All can be tuned per agent.
The default escalation rules
- Positive reply — any reply scored as positive/engaged escalates to a human.
- Scheduling request — "can we book time" type replies escalate to an AE.
- Complaint — frustration, opt-out, or legal-flavored language escalates immediately.
- Pricing pushback — "how much does this cost" or "too expensive" escalates.
- Exec request — "can I speak with your CEO/VP" escalates immediately.
How escalation works
When a trigger fires, the agent pauses on that prospect, drops a card in the Handoff Queue, and notifies the assigned human (via Slack or email). The human reviews the conversation context, decides the next action, and either takes over or returns the prospect to the agent.
Tuning escalation per agent
Agents → [Agent] → Settings → Escalation Rules. Add custom triggers (specific keywords, account-tier triggers, time-based triggers), adjust thresholds (e.g. escalate only on "very positive" replies, not all positive), and adjust routing (who receives the handoff).
What you should and shouldn't escalate
Escalate things humans handle better: real conversations, complex objections, executive interactions, anything legally sensitive. Don't escalate things the agent handles fine: meeting scheduling (use the AI Voice Agent or scheduler), simple FAQ replies, basic enrichment requests.
