Dealing with a very strange bug here with the Bot Analytics view within the Agent Performance dashboard. The environment is configured in the Europe/Paris region, and the issue persists across multiple queues. When examining the Conversation Detail view for interactions routed through the AI Bot flow, the Handle Time metric frequently shows a value of 0s or null for conversations that clearly involved a handoff to a human agent. The Queue Activity dashboard, however, reports accurate Wrap-up Time for these same sessions. This discrepancy creates a significant gap in the Performance Views, making it difficult to validate the Predictive Routing success rates. The specific Architect flow is designed to capture Bot Sentiment before the transfer, yet the downstream metrics in the Performance Dashboard do not reflect this engagement. Is there a known limitation with how Genesys Cloud aggregates Bot interaction data into the Agent Performance reports? The issue appears to be isolated to conversations initiated via Web Chat and does not affect Voice channels. Any clarification on the data retention policy for Bot segments would be appreciated.
“reportingMetric”: “bot_handoff_duration”
Zendesk calculates handle time differently, so GC often shows `0s` until the bot explicitly hands off to a human agent. Verify the `handoff` event is firing in Architect to trigger the correct metric.You might want to check at the conversation/events stream instead of relying on aggregated dashboard metrics. The handle time often stays null until the final disposition is posted, which can lag during high concurrency.
Check if your load test is dropping connections before the handoff event fires. This usually results in incomplete sessions that the analytics engine skips.