video
Service icon

Arcanum Emotion Analysis from Audio/Video

Stable version 1.0.0 (Compatible with OutSystems 11)
Published on 17 May by 
0.0
 (0 ratings)
video

Arcanum Emotion Analysis from Audio/Video

Details
Analyse audio data to gain insights and improve user engagement with AI emotion analysis.
Read more

Emotion Analysis from Audio

Utilise call recordings to improve customer support and uncover valuable insights. 

Benefits

  • Identify patterns in customer emotion - Analyse calls to extract information on customer emotions, utilising this valuable data in order to gain insights for decision making.
  • Monitor Sentiment in Calls - Use call data to see what general emotions are during team and client interactions - allowing you to keep tabs on customer satisfaction and staff wellbeing.
  • Improve Customer Experience - by understanding the emotion of your customers sooner, you are able to keep them happy and intervene to reduce customer churn.


Pain point

Millions of calls and voice messages are being recorded every day, tracking customer interaction and other oral communication. The ability to categorise and contextualise the underlying emotion of those messages adds important insight to the effectiveness of customer support and employee wellness.

Solution

This workflow allows the user to examine the underlying spectrum of emotions within 8 categories from recorded messages. The workflow ingests an audio or video recording and transcribes any conversation. It then analyses this transcription and categorises each sentence to one of eight broad categories: Joy, Trust, Fear, Surprise, Sadness, Disgust, Anger, and Anticipation.

Usage

The input to this model is an mp3 file. Internally, the model transcribes the file and breaks it down into sentences, that are then analysed for underlying emotions. The output of the model is a JSON file with a formatted response of {sentence:emotion} pair.

Expectations/Limitations

This model performs best on emotionally heavy recordings such as reviews. 

The language used in the audio transcription is analysed, but the tone of voice or pitch are not.

 

What’s new (1.0.0)
Reviews (0)