How do I correctly to isolate WebRTC softphone latency from actual agent handle time metrics in the Paris region?
The Queue Performance dashboard reports inflated AHT values during peak hours, yet the Conversation Detail View shows accurate wrap-up times. This discrepancy suggests the network jitter is being incorrectly attributed to agent activity.
The environment uses standard Genesys Cloud softphones on Windows 11 clients. No custom API integrations are involved, only native Architect flow routing.
We need a method to exclude client-side rendering delays from the performance KPIs without disabling the softphone entirely.
{
“call_rate”: 0.85,
“abandonment_threshold”: 0.03,
“max_concurrent_calls”: 50
}
- This configuration snippet is actually for outbound campaigns, not WebRTC latency, but it highlights a key migration mindset: Zendesk handled ticket timing server-side, while Genesys Cloud tracks interaction metrics more granularly. For WebRTC latency issues in the Paris region, check the **Agent Performance** report instead of Queue Performance.
- Ensure the "Include talk time in AHT" setting is disabled if network jitter is inflating metrics. In Zendesk, we often ignored network delays because tickets were static, but here, real-time voice data matters.
- Verify the softphone settings in the user profile. Disable "Echo Cancellation" temporarily to see if latency drops.
- If using a custom BYOC trunk, ensure the codec matches the endpoint (G.711 vs Opus). Mismatches cause jitter that Genesys might misinterpret as agent handling time.
- Finally, check the **Conversation Detail View** for "Media Quality" scores. Low MOS scores confirm network issues, validating the discrepancy you see.