Hey everyone. I am a network engineer trying to track down some really bad voice quality issues for our remote agents. The Quality Management team keeps flagging interactions where the agent sounds robotic, and they want me to fix it. The problem is they just hand me a massive list of interaction IDs. Is there any way to pull the MOS score directly into the Quality Management evaluation view, or map the WebRTC media stats directly to the QM scorecards? Right now, I am manually taking their evaluation IDs and cross-referencing them against the Voice Diagnostics view in a spreadsheet. It is taking forever!
You cannot natively embed the network MOS score directly into the standard Quality Management evaluation forms or the immediate evaluation view. The media diagnostics and the evaluation components are stored in separate datasets. You must use the Platform API to automate this correlation.
You can query the /api/v2/quality/evaluations/query endpoint to retrieve the evaluation scores, and subsequently query the /api/v2/telephony/providers/edges/conversations/{conversationId}/metrics endpoint to retrieve the corresponding MOS and Jitter statistics for the agent’s WebRTC leg.
Greetings! To add some valuable context to the previous reply, the correlation between poor audio quality and low evaluation scores is a highly recognized challenge, especially when managing multiple Bring Your Own Carrier SIP trunks across diverse geographical regions. When you extract the media statistics via the API, it is absolutely critical that you do not just look at the overall conversation MOS! You must traverse the JSON payload and specifically isolate the media segment belonging to the user participant, rather than the external customer leg. A poor MOS on the external leg indicates a problem with the caller’s cell service, whereas a poor MOS on the user leg confirms your suspicion regarding the agent’s remote VPN connection! Building a custom PowerBI dashboard joining these two API datasets is incredibly rewarding!