Genesys Cloud - Main

 View Only

Sign Up

Back to discussions
Expand all | Collapse all

Inconsistent Agent Empathy and Sentiment Scores for Phonecall, Callback, and Transfer Interactions

  • 1.  Inconsistent Agent Empathy and Sentiment Scores for Phonecall, Callback, and Transfer Interactions

    Posted 20 days ago

    Subject: Inconsistent Agent Empathy and Sentiment Scores for Phonecall, Callback, and Transfer Interactions

    Description:
    The Agent Empathy and Customer Sentiment score values returned by the getSpeechandtextanalyticsConversation API are not being retrieved accurately

    Details:

    • API Observation:
      • The score values retrieved through the API are not consistent across different interaction types.
    • Transfer Interactions:
      • For Agent 1, both Agent Empathy and Customer Sentiment scores are blank.
      • For Agent 2, the scores are displayed correctly.
      • In Genesys Fullclient, the Agent Empathy score is missing for both agents.
    • Callback Interactions:
      • The Agent Empathy score is not visible in Genesys Fullclient but appears as 100 in the API response.
    • Additional Observation:
      • When the Agent-Customer conversation duration is very short, the scores appear as 0 in both Genesys Fullclient and the getSpeechandtextanalyticsConversation API results.

    #API/Integrations

    ------------------------------
    Sai Shankar
    x
    ------------------------------


  • 2.  RE: Inconsistent Agent Empathy and Sentiment Scores for Phonecall, Callback, and Transfer Interactions

    Posted 20 days ago

    Hello Sai, 

    I hope I can explain the inconsistencies your seeing. 

    Let me break down each scenario. 

    Lets start with transfer interactions. The blank scores for agent 1, while agent 2 has the correct scores, is expected behavior. During transfer, the system treats segment separately. The missing Agent Empathy score in Genesys could be related to the following factors: insufficient conversation duration for accurate scoring or system treating transfer segments as separate interactions. 

    Next lets move on to the callbacks. The disconnect between Genesys and the API suggests that the interaction might only contain empathetic phrases, which results in a perfect score of 100. But it should still be showing in Genesys this might be something Support will need to look into. 

    Finally for short duration convo's the score of 0 in both Genesys and the API are expected. This is because there may not be enough speech content to analyze and the system requires a minimum amount of conversation data to generate meaningful scores. 

    To wrap this up:

    1. Transfer interactions, what you are seeing is expected
    2. The API is showing 100 while Genesys is showing no data, this requires a support case to see where the disconnect is coming from. 
    3. Short duration convo's, what you are seeing is expected.

    Cheers, 



    ------------------------------
    Cameron
    Online Community Manager/Moderator
    ------------------------------

    Original Message


Related Content