Sentiment Metrics

The Sentiment tab provides easy yet granular visibility into the sentiment analysis and scoring of all customer interactions.

Sentiment analysis uses Artificial Intelligence (AI) and machine learning to analyze call transcripts 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".

The call distribution by sentiment is grouped into 5 categories and marked with distinctive labels:

• Very negative
• Negative
• Neutral
• Positive
• Very positive

Sentiment score vs sentiment label

The sentiment labels are set based on a numeric sentiment score that is calculated for each call, where each label covers 20% of the values on a scale. The MiaRec platform uses a numeric scale from -100 to +100 for a sentiment score.

A relationship between labels and scores is presented in the following list:

• Very negative (scores below -60)
• Negative (scores between -60 and -20)
• Neutral (scores between -20 and +20)
• Positive (scores between +20 to +60)
• Very positive (scores above +60)

The Sentiment page shows the overall distribution of calls by sentiment as well as the chart, which represents daily sentiment trends through the selected period.

Depending on a drill-down level, the Sentiment page also shows a distribution of calls by sentiment for each group or user.

The inline labels show the total number of calls in each sentiment category. By hovering a mouse cursor over a label, you can see the sentiment category name.

Sentiment metrics on a user level

When drilling down to a user level, the sentiment metrics are presented on the Overview page:

Sentiment score in Calls tab

By navigating to the Calls tab, you can view the sentiment score in the corresponding column of the calls list.