POST /api/v2/ai/bot/flows/{flowId}/publish returns HTTP 400 with payload {"message":"Invalid flow state","errors":[{"detail":"Cannot publish while agent availability is being recalculated"}]}. This occurs exclusively when the WFM scheduling module triggers a batch publish event for agents in the America/Chicago timezone. We are using Genesys Cloud v2 with the latest Architect SDK. The issue seems to tie into how the AI bot engine handles concurrent updates to agent profiles during the weekly schedule lock period. Specifically, when an agent’s shift swap is approved and their availability changes from Unavailable to Available mid-publish, the bot flow validation fails. We’ve verified that the API rate limits are not being exceeded, and the flow JSON is valid outside of this window. The error log shows the failure happens within 200ms of the WFM schedule commit, suggesting a race condition between the WFM service and the AI bot publishing service. Is there a documented way to sequence these operations or add a retry mechanism in the SDK to handle this transient state conflict?