Hello. I am a recording export specialist and I am currently working with our quality management team to modernize our evaluation process. We want to use the Genesys Cloud sentiment analysis scores to automatically flag interactions for manual review. Specifically, we want to route any call with a ‘Negative’ customer sentiment score directly to a supervisor’s evaluation queue. How can I build a workflow that automates this assignment based on the final sentiment analysis results?
Hello. This is a fantastic idea! We have implemented something similar for our Japanese team. You can use the Notification API to subscribe to the v2.quality.evaluations topic, but that only triggers after an evaluation is started. For your use case, you should use an EventBridge rule that triggers when a conversation is finished.
The event contains the sentiment data. You can then use a Lambda function to call the Quality API and create an evaluation assignment for the supervisor.
I am a speech analytics manager and I must warn you that relying solely on the ‘Overall Sentiment’ score can be misleading. A customer might start frustrated but end the call very happy. You should instead look at the ‘Sentiment Trend’ metric.
If the trend is consistently negative, then it is a high-priority for review. I suggest you fine-tune your topic detection first to ensure the sentiment is actually related to the agent’s performance and not just a product issue.