Patch Release CX4.10.6

Agents can now place chats on hold when a voice call arrives i.e. when the media gets interrupted. Instead of disconnecting/re-routing, ongoing chats pause while the agent handles the call, with the option to resume during or after. Administrators can configure this behavior in Conversation Studio to either hold or re‑queue chats, aligning with their routing strategy and customer experience goals.

The Bot Framework now processes both CIM and standard Rasa messages. A new standard message parser now ensures that plain and button-based messages are correctly processed in either format, improving compatibility with existing Rasa bots and simplifying integrations.

Issue: Previously, Agent task entries in the advisor queue could reference a CONVERSATION_ID that did not exist in the conversations table. This caused inconsistent data and made troubleshooting difficult.

Resolution: The system now validates that the customer is present in an active conversation before allowing an Agent Transfer. If no valid conversation exists, the transfer request fails, preventing orphaned agent task entries and ensuring data consistency between the advisor queue and conversations table.

Issue: Previously, in some cases, no task disposition was recorded for an agent task even after the agent joined the conversation. This resulted tasks with null dispositions, affecting reporting and operational reviews.

Resolution: The Conversation Manager now ensures that the agent does not join a conversation if it has already been closed. This was a synchronization issue where the agent subscription process was being triggered in parallel with the conversation closing process.

Issue: Previously, the Oldest Queued Time in the Summary Dashboard did not match the value in the Queue Conversation Details view. In addition, the timer was not reset after the oldest chat closed, causing inflated waiting times in dashboards.

Resolution: Real‑time reporting logic now recalculates Oldest Queued Time whenever the oldest queued conversation changes. The Summary Dashboard and Queue Conversation Details display consistent, accurate values, giving supervisors reliable visibility into actual queue wait times.

Issue: In some cases, conversations handled by a bot continued to appear as active in real-time dashboards even after they were closed in the backend. This led to inflated active conversation counts and stale information for supervisors.

Resolution: The real-time reporting logic now reliably updates conversation states for bot-handled interactions, ensuring closed conversations are removed from the dashboards. Active conversation counts and related metrics now accurately reflect the actual live workload.