Release Notes - Hotfix release DC2024.01g

WFM plugin for Verint - CSI pulse report start time improvement.

The Verint CSI/Pulse module expects call summary WFM data to be sent at a 15 minute interval over the past 15 minutes to predict WFM occupancy. At start of business day a 15 minute offset was introduced resulting in a delay of 15 minutes of WFM intelligence for the remaining business day. This mechanism has now been improved (by subtracting report interval start time from CSI time) to reflect data closer to actual occupancy usage.

Audio beep to notify voicemail can be left or when supervisor joins lowered volume.

The amplitude of the audio beep heard when to leave a voicemail, but also used to indicate when a supervisor joins a call has been lowered. the perceived loudness was too high.

Remove additional Reason Codes presence states from SharePoint.

We have removed prematurely added new Reason Code presence states ("Do Not Disturb", "Be Right Back", and "Available") added to the PresenceState column in SharePoint as the WebAgent does not yet have these in its dropdown menu.

Dialogue Studio - Identical EventBus events are now not sent multiple times.

Regardless of how many UCCs are sending an event, if this is an single/identical event (because it's shared by all UCC's) was sent multple times we now send this only once.

WFMPlugin for TeleOpti by Calabrio - Introduce Settings for Unique Agent per UCC identification.

If an Agent is registered as agent in more then one UCC, no easy distinction could be made in TeleOpti for which UCC this agent did specific tasks. We have now added additional PluginSettings to add a unique identifier name to an agents if in mutiple ucc's.

See: Teleopti.RTA.UserCodeSuffix, and Teleopti.HD.AgentIdSuffix and subsequent items marked new in dc2024.01g

WebAgent - Tooltip for iFrame plugin.

A tooltip has been added to iFrames displayed in the webagent.

Dialogue Studio Custom Method in Say Node was not sending audio.

When a 'Say Node' was deployed with 'method' set to 'custom' and any voice from a TTS provider was selected (like 'nl-NL Female'), the playback would not be heard. just silence while the call was kept alive until progressing to a next node step. This has now been resolved and audio playback is actually heard.

Dialogue Studio could crash when importing a built-in Node-Red flow.

The Dialogue Studio instances on the Dialogue Cloud environment have been enhanced to prevent them from crashing when importing a Built-in Node-Red flow.

Dialogue Studio Action Node - Show used config visually.

We now show a visual cue beneath the action node to show which ucc is used. This is updated upon a successful call to the action Node (will also show information if incoming node works).

NodeRedPlugin - Entering a queue crashes UCC.

In very specific scenarios, when an audio call was terminated in an undefined manner while being processed though Dialogue Studio, this could trigger a crash of the UCC service. This InvalidOperationException has now been mitigated and is fixed.

Stop adding to BlockingCollection before Dispose in various integrated Core plugins.

Several Core plugins could raise a BlockingCollection after a shutdown initiation of a UCC service preventing the UCC from shutting down (specifically disruptive when a restart was requested/expected). This defect was also a cause of a small memory leak in these plugins when Consultative transfers were processed.

WFMPlugin - Error updating SkillPerAgents.

The WFM plugin could not process some changes (updates and deletes) in the SkillPerAgent list entries. This has now been addressed.

Dialogue Studio - Token cache not working properly for AnywhereNow Authentication.

When Administrators were working on a Dialogue Studio flow in the DS Canvas they were sometime faced with an unresponsive screen or required re-login to resume the session. The refresh and cache of tokens for sessions has been enhanced and should now be working for the entire day an Administrator may be working on a flow. Important to note, operational dialogue flows were not affected by this.

Dialogue Studio IVR Interactive Voice Response, or IVR, is a telephone application to take orders via telephone keypad or voice through a computer. By choosing menu options the caller receives information, without the intervention of a human operator, or will be forwarded to the appropriate Agent. queue position not correct.

When using Dialogue Studio to play back the actual queue position this was off by one due to an interpretation error. This has now been corrected.

Dialogue Studio repeated early termination of welcome message caused Dialogue Studio container to crash.

During internal development with other dependent Core services and under stress testing it was found a specific scenario (many repeated calls terminating before finish of playing the welcome audio file) would cause the Dialogue Studio container to crash.

Call classification was not working when a direct transfer was made to another agent also using webagent (because AgentSipAddress becomes phonenumber after direct/blind transfer).

When a direct transfer was made to another agent also using WebAgent, (so not a regular other Teams or phone user) we fixed that AgentSipAddress will be the sent to their WebAgent (only if it's an agent), instead of only phonenumber, to make call classifications work for the transferred to agent.

MaxFailedPresenceSubscriptionsPerUri not used with failing presence subscriptions.

Fixed that HandlePresenceSourceSubscriptionFailing event handler in PresenceProvider.cs is fired again when needed. We fixed the way presence detection works for sipaddresses which cannot be found anymore (e.g. agent account discontinued or a typo) to prevent unnecessary retry events.

OutboundSessionAgentFeedback was not corresponding with 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. response.

When an outbound call (from an AnywhereNow Dialpad or a campaign Dialer call) encounters e.g. a busy signal (486 Busy Here) from the end-customer endpoint, the feedback to the Agent (who is already reserved and waiting on the call) will now reflect this error.

PlayPrompt rapid loop when certain IVR questions are empty or misconfigured playlist.

If IVR queue or playlist audio files are empty (misconfiguration) we now do not trigger a play event anymore to prevent excessive memory consumption (which until now played multiple silent sound bytes in rapid succession).

Dialogue Studio queue IVR configuration with fallback skill not triggering the Waiting Queue Node.

We have fixed that Dialogue Studio "Waiting queue" is triggered when a caller is going to a fallback skill of a preferred hunt action. Before this Waiting Queue was not audible, just silence.

Fixed UCC not starting due to incorrect phone number format in forward skill.

We have fixed that UCC service now starts up even if an incorrectly formatted telephone number was entered into the ForwardToSip in (Chat)ForwardSkills fields.

Wrong CDR Call Detail Records (CDR) are the meta data of a converstation: who talked to whom, at which time, for how long and with which identity among other things (skill, optional classification, etc) which gets stored in your SQL (analytical) database. data when hunt failed because of customer leaving.

Fixed CallDetail CDR data when the customer hangs up during the hunt of the first agent when more are reserved to hunt Agents that are not hunted are now not marked a missed call in "Advanced Statistics" in that scenario

Changed the Graph retry policy to get users from graph in case of failure.

Changed the retry policy when failure to get users from graph. For the first 4 retries, it retries exponentially and then it retries each 5 minute forever.

DashboardService - REST API get /{uccname}/agent to include includeFormalStopped.

DashboardService REST API get /{uccname}/agent now includes the includeFormalStopped query parameter.

Added Graph presence polling interval configuration setting.

Configurable parameter PresencePollingInterval is introduced in the config.xml file under the <MsGraphPresenceSource> tag. Value is in the range 1-120 seconds. The default one is 1 second. Note: Changes done here require a UCC service restart to take effect.

Microsoft Azure TTS for "message busy" working with queue playlist flow active.

We have fixed that "message busy" will use the Cloud speech (TTS) synthesizers (Microsoft/Google) when the queue playlist flow is active. It would in some setup variations fall back to the Legacy TTS voices.

Dialogue Studio - Debug Node could restart container.

We had identified a problem in the latest version of the platform underlying Dialogue Studio. When the out-of-the-box Debug node was used in call flows, an excessive amount (compared to earlier version of DS) of data would be contained in the debug body overloading the resources in Dialogue Studio if multiple simultaneous calls were being handled. This would cause restarts or failures of Dialogue Studio. Deleting debug nodes when running in Production was the only solution. We have now contained the amount of data which will be sent by Core into a debug node but we still encourage to Only use Debug nodes during building and testing of your flows. Always remove them afterward for readability of your flow and performance. Please read our page on How to Disable All Debug Nodes at Once for more guidance and prior warnings about memory consumption of Debug nodes.

Dialogue Studio fixed a memory leak in ask-and-wait node.

We patched a memory leak in the ask-and-wait node, resulting in improved memory consumption.

Message Busy continued playing file after agent accepted call after a when a queueplaylist was configured.

In a specific scenario, when a playlist was configured and enabled as QueuePlayList in an IVR, a call was waiting in queue and a QueuBusyTimeout had passed at least once, when an Agent then picked up a hunted call (because the agent became eligible for hunt then) the Playlist would continue playing until the end of file(s) interfering the agent and customer talking. This should no longer be the case.

Dialogue Studio - lonCancel events not logged and treated as error anymore.

We have reduced excessive logging as a result of onCancel events (i.e. when a call gets aborted by a participant midway an IVR or queue message) being treated as error messages but as normal events, improving memory load.

WFM Plugin for TeleOpti by Calabrio - Now recognize agents upon agent creation without the need to assign skill.

We have changed the plugin for TeleOpti by Calabrio to send new agent data over to TeleOpti before a skill has been assigned in AnywhereNow. Improving Agent recognition for better adherence overview.

DialogueManager and WebAgent - MessageReady events are now sent in chunks.

For text based Dialogues picked up by a new agent from a queue with many messages (long asynchronous conversations like e-mails or long chats with multiple replies) the data to be displayed to the new WebAgent is now sent in chunks to optimize memory and graphical display of the message.

WebAgent - List of Reason Codes was not fully scrollable.

When more than 18 Reason Codes were entered the scrollable list in WebAgent did not display all of them and ommitted some in seemingly random method (so not just everything beyond nr. 18) This has now been fixed.

WebAgent - Make calls to Microsoft Graph from WebAgent are optimized.

Queries to Microsoft Graph from WebAgent (i.e. when searching for contact details and displaying in a list) will now be handled in a serial manner and parallel requests are actively detected and refused. Where possible calls are grouped in batch (JSON batching), and retry timeouts are increased over time to reduce load to Graph and throttling Throttling, in the context of a software cloud delivery service, is the practice of temporarily suspending or diminishing access to a service offering to a cloud customer. Because cloud platforms are mostly multi-tenant, shared platforms and the usage at any moment of many customers is unpredictable, throttling is a normal and logical procedure of retaining a cloud service for all customers, sometimes at the expense of one or a handful of customers who are at that time using too many resources for the backend platform. Not only a customer itself, but also a "noisy neighbour" on the same platform can be the cause of throttling. Usually service will be resumed at a normal level after a short cool-down period, but can even take up to 24 hours or more. In Microsoft 365 throttling is usually indicated by "429 errors" or "server too busy" messages in logfiles or replies. by Graph. This should greatly improve the near-realtime representation of contact details retrieved, like Presence state and activity, Contact photo and jobTitles.

WebAgent - Salesforce CTI activity is not being closed by the WebAgent when auto finish is enabled.

CTI end of Activity was not being sent so when "AutoFinishAudioDialogue" set to true. So when activities (like new phone call) being pushed to to Salesforce they remain open (i.e. without closed time) in the platform for that contact or account.

WebAgent - Some callers were not bringing up associated case in CRM Customer Relationship Management, or CRM, is (usually) a software-based, data management method to deal with interactions with customers and potential customers..

When using WebAgent connected with a CRM Service not always were cases collected while these were available. We have now updated queries to improve results.

WebChat - transcript now always downloaded as txt file instead of sometimes opening in other browser tab.

Fixed an issue where the transcript would open in a new browser tab instead of downloading the contents if customer clicked the Download Transcript button. This would only happen if the webchat is embedded in a different domain.

Email Dialogue Provider - Fix for html in unformatted emails processed by EWS (Exchange Server).

A plain text email with content that includes html tags (e.g. <an@email.com>, or <body>) should have that content escaped so it will be displayed as text instead of interpreted as html. This is now fixed also for email being handled by an EWS (Exchange Web Service) or rather from a Microsoft Exchange Server.

WebWallboard with AnywhereNow Authentication enabled disconnected every 8-10 minutes & took 30s to reconnect

We have now improved the way token expiry is handled in the WebWallboard used in this bundle. When AnywhereNow Authentication is enabled for the client applications, the connection to the information stream from which the WebWallboard receives its information now remains active, even over longer times.