Human handoff
Make escalation visible
ZChat treats handoff as a product surface: visitors can ask for help, AI can escalate uncertainty, and operators keep ownership, context, and evidence in view.
Visitor-requested help
Customers can request a human without waiting for AI confidence to fail, and the handoff reason stays attached to the conversation.
- Visible request
- Reason captured
- Conversation updated
AI escalation
AI can move sensitive, uncertain, or policy-bound work to an operator instead of improvising an answer outside approved boundaries.
- Uncertainty
- Sensitive topics
- Approved boundaries
Operator ownership
Assignment and follow-up stay human-owned so support leaders can review who took over, what context was present, and whether the path worked.
- Named owner
- Timeline context
- Review trail
Every escalation should leave a reviewable trail.
| Capability | Status | Detail |
|---|---|---|
| Direct visitor handoff | Available | Widget handoff requests update the conversation and preserve the reason operators need for follow-up. |
| AI uncertainty escalation | Available | AI escalation routes risky or unresolved work to humans instead of presenting unsupported certainty. |
| Attachment and context escalation | Available | Uploaded context can escalate to the dashboard while keeping the conversation record available for operators. |
| Webhook notification | Configured | Handoff events can notify external systems when a workspace has configured an approved webhook. |
| Workforce-management parity | Not available | Not built: workforce-management scheduling, staffing optimization and queue automation. ZChat shows staffing against chat volume; shift scheduling stays in your own tools. |
ZChat puts its effort into a clear handoff to your team, an agent who owns each conversation, and a record of what the AI did.