Observing a persistent divergence in average handle time (AHT) metrics when agents utilize the embedded WebRTC softphone versus standard desktop applications. The EU-West environment registers a 15% increase in post-call work duration for WebRTC sessions, despite identical flow configurations. This latency appears to delay the ‘Call Ended’ state, artificially inflating queue occupancy reports. Is this a known synchronization issue with the browser-based client, or does it require specific flow adjustments to align with standard telephony metrics?