Sentiment

The Sentiment action examines a piece of text, assigns a positive or negative value to the text and returns a sentiment score.

Note: Functionality of this action is text-related only.

For more information about how to use IBM Watson’s Text Analysis, click here.

 

Sentiment differs from Tonality in that provides more simple and easily comparable feedback. Tonality is more nuanced and evaluates different emotions.

 

Inputs

 

Input Name	Description	Default
Text Feed	Choose what you want to record and then analyze for sentiment. You may have other channels available depending on how you have built your flow. For example, you may have already created a recording or received an incoming SMS. ANI: Records the caller text DNIS: Records the receiver text ICALL_#.CHANNEL_ID: Records both parties' text	User Input Required
Cloud Vender	Indicates the vendor the text is sent to for analysis. Choose from IBM, AWS, or Google	IBM Watson
Sentiment	The name of the variable created for the measurement of the sentiment captured.	SENTIMENT_#.sentiment

 

Variables

The Sentiment action creates the following variables to be used and evaluated in the rest of the flow:

Variable Name	Variable Format	Description
SENTIMENT_#.sentiment	positive/negative (not capitalized)	Stores the sentiment value. The number in the name is the Sentiment action's number.

Check out How to Use Variables in SmartFlows for more information about using variables in your flow.

 

Analyzing the Sentiment

After the Sentiment action, you need to specify what you want to happen based on the sentiment’s result. To do so, you can use either a Condition or Switch action. However, we recommend the Switch action since it allows you to be more specific and have more control over what you do next.

 

Example Configuration

The following is an example of how you could evaluate the sentiment with either the Switch or Condition action.

Variable Name	Condition	Value	Description
SENTIMENT_#.sentiment	contains	positive	Can be used when you want to determine if the customer left the interaction satisfied.