Stuck on reconciling the Service Level metrics in the Performance Dashboard when AI bot interactions are introduced into the flow logic.
Our EU-West BYOC environment is currently exhibiting a significant divergence between the actual agent engagement times and the Service Level percentages reported in the standard Performance views. The issue appears to stem from the specific configuration of the AI Bot handoff within the Architect flow. When a customer interacts with the bot for an extended period-specifically around 45 seconds to resolve initial queries before transferring to a human agent-the dashboard continues to count this duration against the Service Level target. This results in an artificial inflation of the wait time metric, suggesting that agents are missing their 80/20 targets when, in reality, the customer was being assisted by the automation layer.
The Architect flow is configured to transfer the interaction to the primary support queue only after the bot determines it cannot resolve the issue. However, the Performance Dashboard does not seem to distinguish between the ‘bot processing time’ and the ‘queue wait time’ for the purpose of the Service Level calculation. Consequently, the metric reflects the total elapsed time from the initial call arrival to the agent answer, rather than just the time spent in the queue. This discrepancy is causing significant confusion for our workforce management team, as they are attempting to optimize agent staffing based on data that includes non-agent resolution time.
We have verified that the conversation detail views correctly show the bot interaction segment, but the aggregated dashboard metrics remain skewed. Is there a specific configuration within the Architect flow or the Dashboard widget settings that allows for the exclusion of bot interaction time from the Service Level calculation? We require the metric to reflect only the time the customer spent waiting in the queue for a human agent, excluding the automated resolution attempt. Any guidance on aligning these metrics with business expectations would be appreciated.