Ran into a weird issue today with the Agent Scripting feature in our EU-West BYOC environment. We are utilizing Architect flows to enforce compliance scripts before allowing agents to transfer calls. The flow logic triggers a Play Prompt action, followed by a Wait For Input node. However, when an agent attempts to transfer a conversation after the script completes, the system fails to recognize the transfer intent. Instead, the agent is locked into the current interaction state, and the transfer capability becomes unresponsive for approximately 30 seconds.
This delay is unacceptable for our KPIs, particularly regarding Average Handle Time (AHT). The Architect flow shows no error logs, yet the agent experience is severely degraded. We have verified that the script duration is under the 10-second threshold defined in our policy. Is there a known latency issue with the scripting engine in the EU-West region, or is this related to a configuration mismatch in the flow’s post-interaction actions? We need to ensure the transfer action is decoupled from the scripting validation step.