Quick question about Predictive Routing configuration within our EU-West BYOC environment.
Background
Our organization utilizes Predictive Routing to optimize agent utilization based on historical performance metrics. The current Architect flow routes inbound contacts to a specific queue where the routing strategy is set to ‘Predictive’. The goal is to match callers with agents who have a higher probability of successful first-contact resolution based on past interactions.
Issue
The dashboard indicates that the Predictive Routing model is active; however, the assignment behavior appears inconsistent. During peak hours, the system frequently defaults to standard FIFO routing despite the Predictive strategy being enabled. This results in a noticeable drop in first-call resolution rates, which contradicts the expected performance gains outlined in the initial deployment plan. The discrepancy is most evident in the Agent Performance views, where predicted match scores do not align with actual contact outcomes.
Troubleshooting
- Verified that the Predictive Routing toggle is enabled at the organization level.
- Confirmed that sufficient historical data exists for the model to train effectively.
- Checked the Architect flow to ensure no manual overrides are forcing standard routing.
- Reviewed the queue settings to confirm no conflicting routing priorities exist.
Has anyone encountered similar inconsistencies where the Predictive model fails to influence routing decisions during high-volume periods?