Quick question about the interaction between WFM schedule publishing and Architect flow logic. We have a custom IVR flow designed to route calls based on real-time agent availability for our Chicago support team. The flow uses the ‘Get Queue’ task to check the number of available agents in the ‘L1 Support’ queue. If the count is zero, it routes to voicemail; otherwise, it adds the caller to the queue.
The issue arises immediately after an agent completes a shift swap in the WFM portal. The swap is approved, and the schedule is published successfully via the API. However, the Architect flow continues to report that agents are available in the ‘L1 Support’ queue, even though the swapped-off agent is no longer scheduled and should technically be offline or unavailable according to the new schedule.
I have verified that the schedule publish endpoint returns a 200 OK status. The delay seems to be in the synchronization between the WFM module and the Architect flow’s view of the queue. It takes approximately 15-20 minutes for the ‘Get Queue’ task to reflect the correct agent count after a shift swap is finalized.
Here is the relevant snippet from the Architect flow debug log:
{
"task": "Get Queue",
"queueId": "L1_Support_Chicago",
"availableAgents": 5,
"expectedAgents": 3,
"timestamp": "2023-10-27T14:30:00Z"
}
The expected count is 3, but the flow sees 5. This discrepancy causes calls to be routed to agents who are no longer on the clock, leading to dropped calls and poor CX metrics. Is there a known latency issue with the schedule-to-queue sync? Should we implement a polling mechanism in the flow to wait for the schedule to fully propagate, or is there a specific API endpoint we can call to force a refresh of the queue status after a schedule publish event?