A customer waiting for your team is no longer left in silence
When the AI agent hands a chat to a person, it can now send one short reminder that the message is still with your team — inside your business hours only.
When the AI agent cannot answer something, it hands the chat to your team and stops talking. That silence is deliberate: two voices in one conversation is worse than one, and the agent staying quiet is what keeps it from contradicting the colleague who is about to reply.
What was missing was an answer for nobody comes. If the handover happened at 7pm, or during a busy hour, the customer could write "hello?" into a chat that nothing would ever answer until someone opened the inbox.
One reminder, and only one
The agent may now send a single short message saying the question is still with your team. It deliberately says nothing about the question itself — that is exactly what stops it from guessing ahead of the person who will actually reply.
It is throttled, so a customer who sends five messages gets one reminder, not five. You set the gap in AI agent settings, and setting it to zero turns the reminder off completely.
Inside your working hours
A reminder at 3am is worse than silence. The reminder is sent only inside the business hours you set, so an evening handover waits for the morning.
Why it does not just give the chat back to the bot
That was the obvious fix and it is the wrong one. Look at why a handover happens: the agent did not know, was not confident, the customer asked for a person, or the conversation had gone round too many times. In every one of those it handed over because it could not answer. Give the chat back and it hands over again, and the customer has been brushed off twice. So the chat stays with your team, and the reminder is all the agent does.