Sentiment analysis is a machine learning tool that analyzes texts for polarity. Polarity refers to the overall sentiment conveyed by a particular text, phrase or word. This polarity can be expressed as a numerical rating known as a “sentiment score”. For example, this score can be a number between -100 and 100 with 0 representing neutral sentiment. By training machine learning tools, machines automatically learn how to detect sentiment without human input and score configuration.
MiaRec Voice Analytics expresses customer sentiment in two different ways: through a numerical score with its associated emoji and through visualization of colors within the keyword list and transcript.
The MiaRec speech engine analyzes identified keywords for positive or negative customer sentiment. Each keyword can be assigned a numeric value, either positive, negative, or zero (
+100). A total sentiment score value is a summary of all spotted keywords' scores. Depending on the number of times a positive or negative keyword is mentioned either by the customer or the agent, a customer score, an agent score, and the total sentiment score (the average of the two) is automatically tabulated and symbolized by the appropriate emoji.
Visualization of the keywords and transcript
Positive keywords, such as "thank you" or "this is helpful", are color-coded in green in the transcript as well as in the keyword list (or any other positive color you like), while negative keywords, like "upset", "angry", or "I expected more", are in shown in red.
This allows you to see at a glance how a call is trending. If a customer is angry at first, does the agent resolve the problem quickly or does it escalate? With MiaRec's visual Customer Sentiment, you can immediately see it without reading the entire transcript, saving you valuable time.
By hovering over the coloured text, you can preview the sentiment score of a specific keyword.