Escalate anything the AI cannot confidently answer, instantly
The value of an AI agent is not how much it answers — it is how reliably it knows when to stop and pass the conversation to a person.
The problem
Dead-end bots that will not let anyone through
Most frustration with support bots comes from the same place: the visitor already knows the bot cannot help, and there is no way out. Loops, restated non-answers and hidden contact routes cost more goodwill than the automation ever saved.
- Visitors get looped through answers that miss the question
- Asking for a person is ignored or buried
- Context is lost when a conversation finally reaches a human
How the platform helps
Handoff is a first-class part of every conversation
Escalation triggers on low-confidence retrieval combined with model uncertainty, on an explicit request to speak to someone, and on confident refusals where the answer simply is not in your content. The thread moves to your inbox with everything already said, and once an agent replies the AI stops generating in that conversation for good.
- Escalation on weak matches paired with model uncertainty
- Immediate handover when a visitor asks for a person
- Full conversation history carried into the shared inbox
- AI generation stops permanently once an agent takes over
What this looks like in practice
One conversation, from an automated answer to a human reply.
- 1
AI answers what it can
The agent replies from your knowledge base while the questions stay within your published content.
- 2
Confidence drops
A question falls outside your material, or the visitor asks to speak to someone.
- 3
Thread is escalated
The conversation is flagged in the shared inbox with the full history rather than answered with a guess.
- 4
Agent takes over
Your agent replies in the same thread and the AI stays silent in that conversation from then on.
Automation your customers do not resent
Start free with unlimited human live chat, and turn the AI on knowing there is always a way through to a person.