Tone Analyzer

Use the tone analyzer node to direct the call flow based on the perceived emotions of the caller. For example, you may want to transfer the call to an operator if the caller seems angry.

The Tone Analyzer node is built upon the IBM Watson Tone Analyzer. See Tone analytics (Classifications).

Call Flow

Prior to the tone analyzer node, save the caller's response to a variable.

Following the tone analyzer node, use a Decision node to direct the caller according to their perceived emotions.

Name

Give the node a name.

Select Text

Typically, a variable contains the caller's response. Type two curly brackets {{ to select from the variables in your account.

For testing purposes, type some text.

Example:

I am happy with the result.

Music on Hold

The caller hears the uploaded audio file while they are placed on hold. Click to upload a WAV file (maximum 3MB).

Response

Match JSON name-value pairs of interest.

  1. Click Preview Response to see the JSON name-value pairs on the Data tab.

    Returned Data

    Key

    Value

    Emotions

    anger

    fear

    joy

    sadness

    analytical

    confident

    tentative

    sympathetic

    polite

    impolite

    The confidence score for each emotion.

    The confidence score is Nil or a number between 0 and 1.

    For example, "anger":0.008622 perceives anger with a confidence score of 0.008622 (0.8622%).

    Top Emotion

    top_emotion

    The emotion with the highest confidence score.

    Top Emotion Score

    top_emotion_score

    The highest confidence score.

  2. Select a JSON name-value pair.

    The path to the selected name-value pair is displayed in the JSON Path field.

    The confidence score for the selected name-value pair is displayed under the JSON Path field.

  3. Give a name to the JSON name-value pair in the Assign path to variable field.

  4. Click Assign.

    The JSON path and variable are added to the table of return values.

  5. Repeat. Select another JSON name-value pair.