Stuck on reconciling WFM forecast accuracy metrics with actual queue performance in our EU-West BYOC environment. The divergence appears specifically when analyzing flows that utilize complex routing logic within Architect. While the WFM module projects a 95% service level based on historical shrinkage and arrival rates, the real-time Performance Dashboard shows a persistent 12% shortfall during peak European business hours (09:00-11:00 CET).
The discrepancy is not accounted for by standard agent adherence issues. Agent adherence reports remain above 98%. The gap seems to correlate with the execution time of specific Architect flow steps, particularly those involving external API calls for customer data enrichment before queue entry. These steps introduce latency that WFM does not appear to factor into the service level calculations.
Is there a configuration parameter to adjust for flow-induced latency in WFM forecasting? Below is the relevant segment of our routing configuration:
flow_id: FLOW_EU_042
routing_strategy: longest_idle_agent
pre_queue_actions:
- type: api_call
endpoint: /customer/profile
timeout_ms: 2500
on_failure: route_to_queue
- type: set_attributes
key: priority
value: high
The timeout is set to 2500ms, which aligns with vendor SLAs, yet the aggregate delay impacts the overall service level. Guidance on aligning WFM projections with Architect flow realities is required.