Beyond Metrics Intelligent Classifiers Transforming CQD-QER and More

This session introduces Intelligent Media Classifiers in CQD QER v5, which use probabilistic machine learning models across network, compute, device, and media modality signals to determine whether a user’s media experience was impacted—moving beyond traditional static threshold‑based metrics that rely on average values of limited network telemetry. These classifiers provide a more holistic view of user experience by analyzing multiple contributing factors and are integrated across CQD reporting surfaces to help administrators detect small or intermittent quality issues and better understand potential causal contributors. 

We also introduce assisted troubleshooting capabilities such as Assisted Drill‑through analysis reports, Silent Test Calls for proactive simulation of Teams audio and video quality, and Admin‑Initiated Remote Log Collection.

Together, these capabilities enable administrators to quickly identify tenant‑level problems affecting media quality across meetings, calls, and events, and collect additional diagnostic data to support post‑incident analysis and remediation workflows.