Good morning, I would need to end in different ways the calls "ended by the user" and the calls "ended by the assistant" so as to create automatic sub-outcomes. For example, if an IVR is detected, the end-end reason should be "ended by IVR", if the call had a positive outcome "POSITIVE", if the customer is not interested "Not Interested", if the customer asks to be called back at another time "Does not have time". Is it possible to further segment the outcomes from the system? Best regards Mattia

@Mattia Good morning, Mattia! Absolutely, it is possible to segment the outcomes of calls based on the specific scenarios you've described. You can customize the end-of-call report to include different

endedReason

values that correspond to the various outcomes of the calls, such as "ended by IVR", "POSITIVE", "Not Interested", and "Does not have time". To implement this, you would need to set up logic within your voice application to detect these scenarios and then send the appropriate

endedReason

in the POST request to the Server URL when the call ends. For instance: - When an IVR is detected, the application could set the

endedReason

to "ended by IVR". - If the call had a positive outcome, the

endedReason

could be set to "POSITIVE". - If the customer expresses disinterest, the

endedReason

could be "Not Interested". - If the customer asks to be called back at another time, the

endedReason

could be "Does not have time". This way, you can track the outcomes of the calls more granularly and take appropriate actions based on the specific sub-outcomes. If you need assistance with setting up this logic or have any further questions, please feel free to ask!

@User I am not clear on how to implement this. Can you give me some clarification on this? Thanks for your support

@Mattia Yes, it’s definitely possible to segment call outcomes further based on how and why the call ended. 1. Start by creating a list of all possible reasons a call could end. 2. Use Tags or Flags to Classify Outcomes 3. Automate Outcome Assignment using LLM 4. Save Sub-Outcomes in Your Database 5. Once you have these sub-outcomes stored, you can analyze them for insights: Example Flow: 1. LLM detects call ended reason intent using transcript and summary: - Customer says: “I don’t have time.” → Outcome = CALL_BACK_LATER. - Assistant detects an IVR → Outcome = ENDED_BY_IVR. 2. Call ends, and the system assigns the sub-outcome. 3. Outcome is stored in your database for tracking and reporting.