Analytics-based text routing allows you to automatically route incoming text interactions based on dataset, sentiment
Overall mood or result of the interaction as determined by analysis of words, phrases, and context of the transcript. Interactions can be positive (blue), negative (red), mixed (dark gray), or neutral (light gray)., frustration, language, priority, or any combination of these factors. The Snippet Studio action can be configured with a script that sends text to the NICE CXone Analytics Based Text Routing service, where the content of the text is analyzed and annotated, and then routed to an appropriate agent based on the information provided by the service.
The types of routing supported are:
- Term routing — Analyzes content to determine the dataset being discussed and routes according to your company specific terms.
- Category routing — Analyzes content to determine the dataset being discussed and routes according to your custom category rules.
- Priority routing — Analyzes content to classify an interaction as “priority” or “non-priority”.
- Overall sentiment routing — Uses sentiment analysis to classify an interaction into four cases: negative, mixed, neutral, and positive.
- End of interaction sentiment routing — Uses sentiment at the end of an interaction to classify it into four cases: negative, mixed, neutral, and positive.
- Frustration routing — Uses frustration analysis on the contact side to classify an interaction as "Frustrated".
- Language ID routing — Analyzes and identifies the language of the text.
- Combination routing — Routes on any combination of the other factors.
Term Routing
Interactions can be routed based on the actual product or service mentioned in the text to an agent with the appropriate skill to handle the interaction.
The routing service analyzes the incoming text, identifies the dataset discussed, and returns an annotation of the dataset, such as “Billing Inquiry”.
Category Routing
Interactions can be routed according to your custom category rules. The category rules can use the custom Company Profile terms or they can use regular keywords. The routing service analyzes the incoming text, identifies keywords, and routes the interaction to an agent with the appropriate skill set.
Priority Routing
Interactions from extremely dissatisfied customers can be flagged for priority routing to specially skilled or more senior agents for handling.
The service analyzes the text and uses criteria like extremely negative sentiment, requests for escalation, risk factors (threats to sue, other items that indicate hazard), cursing, and so on, to assign a contact priority.
Overall Sentiment Routing
You can send positive interactions, which are easier to handle, to new or inexperienced agents, and send more sensitive interactions to experienced agents.
The service analyzes the interaction, identifies the prevailing overall sentiment for the document, and returns an annotation such as “positive”, “negative”, “neutral”, or “mixed”, which can then be used for routing.
End-Of-Interaction Sentiment Routing
You can send positive interactions, which are easier to handle, to new or inexperienced agents, and send more sensitive interactions to experienced agents.
The service analyzes the interaction, identifies the sentiment detected for the end of the document, and returns an annotation such as “positive”, “negative”, “neutral”, or “mixed”, which can then be used for routing.
Frustration Routing
Interactions from frustrated contacts can be marked as frustrated so you can route them to agents who can handle those contacts best.
The service analyzes the text on the contact's side for language that indicates frustration and not simply negative sentiment. When frustrated language is detected, the interaction is flagged as "frustrated".
Language ID routing
Interactions can be routed to an agent with an appropriate language skill.
The service can analyze the interaction and identify the text language, which can then be used for routing.