Operator assistants help operators have fast and meaningful conversations by providing them recommendations on how to proceed with the engagement. The type of recommendation depends on whom the operator assistant is designed to listen:
- the visitor, or
- the operator.
The recommendation can be:
- a response to the visitor’s question that the operator can approve and send (reacting to the visitor);
- prompt on what would be the proper next action to take, e.g. offer CoBrowsing (reacting to the visitor);
- an attempt to coach the operator regarding the language used or products recommended (reacting to the operator).
The operator assistant’s messages are only visible to the operator, not to the visitor. They are also saved to the chat transcript.
According to whose messages the operator assistant is configured to react to, the chat messages of either the visitor or the operator are sent to the underlying engine. Depending on your engine’s logic, accepting and rejecting the recommendations can act as a training tool for the engine, thus the assistance improves the more you use it.