Release Notes - Hotfix Release DC2025.01c

Forward Action node causing log error when retrieving skill.

A fix has been implemented for an issue where the Forward Action node could generate repeated “Could not retrieve skill for SessionQueue Temporary forward skill” log errors when calls were routed through a temporary forward skill. The problem occurred in scenarios where SLA Scheduling attempted to retrieve SLA‑related scoring information for a forward skill that had no applicable SLA configuration. This resulted in unnecessary log noise but did not affect actual call routing behavior. The system now safely ignores forward or null skills during SLA evaluation to prevent the error from occurring.

Corrupted Welcome Message Fix.

This update resolves an issue where Welcome Message (WM) audio files could become corrupted or zero‑sized due to multiple cache‑handling defects. The fix addresses a race condition during file deletion, excessive accumulation of outdated TTS cache entries, unsafe cleanup routines running while audio files were still in use, and a timing gap where valid files were briefly misidentified as unregistered. Together, these improvements ensure that WM audio files remain intact and that cache behavior is stable and predictable across the service.

Dialogue Studio Transcriptor Node Culture Handling Fix.

We fixed an issue where the Dialogue StudioTranscriptor node could fail with a TypeError when a flow included a msg.payload.culture value, preventing the transcription process from running correctly. The node now reliably accepts culture overrides provided in incoming messages. This improves stability for customers who dynamically set culture preferences in their workflows.

Recordings now consistently uploaded to SharePoint.

We have resolved an issue where call recordings were occasionally not uploaded to SharePoint, particularly in high‑volume environments where stopping a recording took longer than 30 seconds. The fix ensures that SharePoint upload scheduling now occurs reliably, even when cleanup operations are delayed. This improves recording consistency and supports compliance requirements for all affected customers.

Correct SIP The Session Initiation Protocol, or SIP, is a protocol for multimedia communication (audio, video and data communication). SIP is also used for Voice over IP (VoIP). SIP has interactions with other Internet protocols such as HTTP and SMTP. Address Handling in WebAgentService.

We resolved an issue where the WebAgentService incorrectly sent a user’s UPN In Windows Active Directory, a User Principal Name (UPN) is the name of a system user in an email address 'like' format. A UPN (for example: john.doe@domain.com) consists of the user name (logon name), separator (the @ symbol), and domain name (UPN suffix). to the DialogueManager instead of their SIP address, which caused active conversations not to be retrieved for agents whose SIP and UPN differed. This fix ensures the correct SIP address is always used when communicating with the DialogueManager. As a result, agents with non‑matching SIP and UPN values will now see their conversations consistently and reliably.

Web Agent & ACDC Login Failure After Upgrade to DC2025.01a.

We have resolved an issue where customers using multiple tenants or SIP domains were unable to log in to the WebAgent and ACDC client after upgrading to DC2025.01a. The underlying cause—support for only a single issuer URL—has been corrected by introducing support for multiple issuer URLs through the new IssuerUrls array in the OIDC configuration. This fix ensures that environments with multiple tenants continue to function seamlessly after upgrading.

EmailDialogueProvider Stops Polling EWS Mailboxes.

We resolved an issue where the EmailDialogueProvider stopped polling EWS mailboxes after upgrading to bundle DC2025.01a. The fix is now included in bundle DC2025.01b, restoring normal mailbox polling behavior without requiring additional plugin settings. Customers upgrading to this bundle should now experience uninterrupted email polling as expected.

Improved File Upload Handling in WebChat.

We've resolved a bug in WebChat where uploading multiple files with identical names would cause all but one copy to become "stuck" in the draft area, making them impossible to remove without refreshing the chat. Now, when customers upload multiple files with the same name and send the message, all copies will be successfully sent, and the draft area will be correctly cleared of all uploaded files, ensuring a smoother and more reliable file sharing experience.

AFAS CRM Customer Relationship Management, or CRM, is (usually) a software-based, data management method to deal with interactions with customers and potential customers. PhoneNumberFormat Pattern Fix.

We resolved an issue where the PhoneNumberFormat “Pattern” setting in the AFAS CRM connector did not correctly interpret configured patterns, causing contact lookups to fail for valid numbers such as 06-31234567 and 063-1234567. This fix ensures that all documented pattern formats are now properly respected, enabling accurate contact retrieval.

CTI Settings Not Updating in Stored Database.

We resolved an issue where CTI‑related Web Agent settings were not being saved or updated in the SQL database, even though non‑CTI settings were correctly persisted. This fix ensures that all Web Agent settings (including CTI options such as Click‑to‑Dial) are now consistently stored and updated as intended. As a result, customers can rely on complete and accurate central settings preservation across their Web Agents.

WebAgent – Primary Agent Selection Fix.

We resolved an issue where selecting a new primary agent during a conference call could fail when the system incorrectly used a SIP address instead of a telephone URI for the transfer request. This fix ensures consistent and reliable primary‑agent handover, even when agents are added through search rather than the frequent‑transfers list.

Web Agent Transfer Pop‑Up Now Dismisses Automatically.

We resolved an issue where the Web Agent transfer pop‑up remained visible even after a consult or delight transfer was successfully accepted, preventing agents from completing the process smoothly. With this fix, the pop‑up now correctly closes as soon as the transferred participant connects. This improvement ensures a more seamless and efficient transfer experience for agents.

Fix for Oversized WhatsApp Message Titles.

We resolved an issue where WhatsApp conversation titles displayed the entire first message, including line breaks, causing the title to overflow and become difficult to read. The title is now properly truncated to a single line with an ellipsis, with the full content available via tooltip for clarity. This improves readability and ensures a consistent and user‑friendly interface.

Correct Presence Display for Searched Users in Web Agent.

We resolved an issue where the Web Agent displayed “Busy” for contacts who were actually “In a Call” when searched, due to the application reading the availability token instead of the activity token. With this fix, the Web Agent now correctly reflects the “In a Call” status, ensuring clearer and more accurate presence information for agents. This improvement enhances reliability when searching and monitoring user presence within the interface.

Fix for Call History Retrieval Error in WebAgent.

We resolved an issue that caused the WebAgent to display the error “Error retrieving callhistory of type All. Unexpected response from call to WebAgent service: Failed to fetch.” This fix ensures that call history is now retrieved reliably without triggering unexpected errors. Customers should experience consistent timeline and call history loading across all affected areas.

Correct Wrap‑Up Timing for Agents Across UCCs.

We resolved an issue where agent wrap‑up values could display incorrect start times, durations, or remaining wrap‑up time when the agent was active in multiple UCCs. The system now consistently updates all relevant agent <-> UCC combinations, ensuring accurate wrap‑up calculations in every scenario. This improvement results in reliable and predictable wrap‑up behavior for all agents.

Improved SLA Scheduling to Prevent Stalled Contact Distribution.

We resolved an issue where contacts queued under SLA-based scheduling could become stuck and never be offered to agents, leading to idle agents despite active queues. The updated scheduling logic now ensures stable queue handling and consistent real‑time selection, even under high load. Queue positions shown on dashboards and announcements now refresh reliably on a fixed interval, improving clarity for both agents and supervisors.

Read more

Affected channels: Voice, Email, Chat

What changed:

Contacts queued under SLA-based scheduling could become stuck and never be offered to an agent, causing agents to appear idle despite a non-empty queue. This was caused by the queue's internal ordering becoming inconsistent as SLA scores changed over time.

The queue implementation for SLA-scheduled UCCs has been redesigned. Contacts are held in a stable list, and reported queue positions — shown on wallboards, dashboards, and heard in position announcements — are refreshed on a fixed 5-second interval. Actual contact selection remains real-time: at every hunt, the system scores all waiting contacts against the current time and picks the highest-urgency contact at that moment. A contact can become the most urgent between two refresh cycles and will still be offered next — it may just not yet show as position #1 on the wallboard.

Additional improvements:

• Queue positions reported to dashboards, wallboards, and position announcements are now stable and refresh every 5 seconds. Actual contact selection by the system remains real-time and is unaffected by this interval.

• Changing the scheduling algorithm on a live UCC (e.g. from First-In-First-Out to SLA) now safely migrates all contacts in queue to the new queue structure, preserving their state and continuity.

• Reduced CPU overhead for UCCs with high queue volumes under SLA scheduling.

Transcription Failure After Teams Internal Transfer.

We resolved an issue where call transcription did not start after an internal transfer performed via Microsoft Teams when the transferred agent was represented with a sip.pstnhub.com URI. The system now correctly handles these SIP address formats, preventing the previous null‑reference error and ensuring transcription reliably starts on the transferred call leg. This improvement restores full transcription functionality for internal transfer scenarios.

Improved Stability for Transcription Services.

We resolved an issue where the transcription service could crash due to an unhandled exception occurring during audio stream conversion. The service now properly handles these exceptions, preventing unexpected shutdowns and ensuring more reliable transcription performance. This improvement enhances overall stability for customers using speech-to-text functionality.

Microsoft Transcription Now Respects Disabled Transcript Logging.

We resolved an issue where the UCC could continue logging speech results even when Transcript logging was disabled while using Microsoft Transcription. As part of this improvement, the Microsoft Cognitive Services transcriptor now uses the newer audio sampling method. This reduces unnecessary logging and provides more reliable transcription processing.