Intents Fallback
Last updated
Last updated
Fallback Intent analytic helps the supervisor view the list of unexpected utterances, or when a customer says something that doesn’t map to any intents in bot skill (every time the bot is not able to match user input with any of the specified intents). As you design bot skills, you identify and define a set of utterances and intents to capture the many possible variations that a customer can speak. When your skill is active, 3Dolphins will try to match all customer utterances to one of the skill’s intents. There may be instances when the customer's request is not appropriate for bot skill. If 3Dolphins maps one of these requests to your intents, it may provide an unexpected response.
First, go to the Analytic menu and click the Virtual Agent menu, then click Bot Overview menu which you can find on the left sidebar. Select Bot to display the intents fallback data.
Component Explanation:
Domain Fallback
The line graph indicates how many bot response failures were based on the FAQ Knowledge for the past week.
Domain Suggestion
The line graph represents the total messages responded to by the bot using the suggestion.
Fallback Rate
The bar graph indicates total messages experienced fallback compared to all messages sent by customers. Meanwhile, the line graph indicates the percentage of total messages experienced fallback compared to all sent messages by customers.
Dialog Fallback
The line graph indicates how many bot response failure based on the Dialog for the past week.
Intention type
The pie graph indicates how many bot response failures based on intention type.
Table Intents Fallback
This table displays all words that are not yet known by bots based on the:
Bot: Bots that are used when customers are interacting, and a fallback occurs.
Intent type: This type of intention consists of the "Domain" and "Dialog". The domain is a fallback message that occurs when a customer is interacting with a bot in the Knowledge FAQ while Dialog is a fallback message that occurs when a customer is interacting with a bot in the Dialog.
Channel: The channel used by customers to interact with bots.
User: The name of the customer who sent the message.
Fallback intent: Fallback intent consists of five types, namely default, request, inquiry, complaint, and suggestion. When customers interact with a bot, and a fallback occurs, that message will automatically sort into five intent fallback types. Each classification is described in detail below: Default: Represents customer messages not identified by the paragraph classifier.
Request: Represents customer messages identified as requests.
Inquiry: Represents customer messages identified as asking questions.
Complaint: Represents customer messages identified as complaints.
Suggestion: Customer messages are recognized as suggestions because they have several similar knowledge bots. For example, when a customer sends a message in the form of 'request' such as 'what excess chatbot 3dolphins' or 'company email' and a fallback occurs, the bot will detect that the message entered into the fallback request. When the intent type is "Dialog", the fallback intent value is a unique id. This unique ID represents the dialog ID that is running when the customer is interacting with the Bot. For example, When a customer is interacting with a bot in the Covid-19 dialog, and a fallback occurs, the fallback intent value in the table will display the Covid-19 dialog ID.
Messages: Messages from customers in the event of a fallback.
