How AI, triggers and operators work together
Three automation layers: who answers first, who stays silent, and when a human steps in.
A customer in TicketON has three possible “counterparts”: triggers, the AI agent and live operators. They don't clash — each has its own zone. This article is about how they split a conversation.
Triggers: before the dialog
A trigger is a proactive widget message: “need help choosing?” on a product page, a promo code for a leaving visitor. Triggers work before the customer wrote: their job is to start the conversation. Anti-spam constraints keep the widget polite (don't show if the chat is open or the customer is already talking to an operator).
The moment the customer replies, it becomes a regular ticket — and the next layers take over.
The AI agent: first line
Who answers a new ticket is decided in AI → Handoff:
- AI answers first — the agent runs the dialog and hands over by the rules below.
- Operator answers first — while the team is online the AI stays silent and only steps in when everyone's offline.
Plus the global mode in Configuration: the AI works “Always” or “When offline”, and only in channels where you enabled it.
Handoff to a human
The AI hands the dialog over when:
- the customer explicitly asks for a person;
- the customer is clearly unhappy (negative messages);
- the question is outside the knowledge base or needs an action the agent has no rights for.
To the operator it looks like a normal new ticket with the whole bot conversation attached. After sorting it out you can hit “Return to AI” — the agent continues with the customer.
The customer's path, summed up
- A trigger starts the conversation (or the customer writes).
- The AI answers first — if enabled and set to answer first.
- Handoff — a human joins on request, negativity or a knowledge gap.
- The operator solves it and closes the ticket — or returns the dialog to the AI.
Didn't find an answer?
Message us in chat — the same team that builds the product replies. Or email [email protected].