Context Overview
Last updated
Last updated
Context is a set of situations that occur during natural communication. Virtual Agent provides Chatbot / Voicebot training feature that learns according to the communication context helping it to give appropriate responses.
If we only consider the statement "What is the interest rate", it is difficult to know whether the customer wants to ask about the loan interest rate or the savings interest rate. However, if placed in the context of "consulting about opening a savings account", the intention is understood as "the customer wants to ask about the savings interest rate".
Using context to establish a response scenario helps the VA manage the information that appears in the context. Even if the customer changes the information, the VA can still identify and extract that information.
Setting context includes input and output contexts.
The output context controls the operation context. When an intent is created, an output context is activated for this intent.
Input context controls the context in which an intent appears in the conversation flow. An intent will be recognized in the conversation flow if all input contexts to which it is specified are currently active.
The Lifespan value is set for the output context. It indicates how many times the customer asks within the context. When a context ends, VA will continue to respond to the active context. The trainer sets this parameter to maintain the conversation context depending on the scenario.
For example: Consider the "Online Savings" product consulting scenario below
Output context: ask_open_saving_account_context sets the lifespan parameter to 5.
Output context: ask_open_online_saving_account_context sets the lifespan parameter to 2.
