The workflow object User Search now supports searching by the LDAP User Name. Also a new return variable is provided which contains the LDAP User Name of found users:

• <prefix>.usersldapauthenticationuser

See the following page: https://wiki.jtel.de/x/SAAF for more details.

The following variable must be set in a Variable Assignment object anywhere in the call flow prior to the Save Recordings object used to save the recording:

• $transcribe_recording

Set this variable to 1 to enable transcription.

The following parameters must be set under Client Master Data … Parameters:

Note, the parameter VoiceMails.Transcribe.Active need not be set. This governs the transcription of Voice Mails recorded within the ACD / Group Rules.

Recordings must be made with the Recording IVR object (see https://wiki.jtel.de/x/OAAF ) and saved (see https://wiki.jtel.de/x/OwAF ) using one of the following methods:

• Portal (which will create a Voice Mail entry in Messaging … Voice Mails)
• SMTP (which will send the voice mail as an E-Mail)
• Distribute via ACD (which will create a Voice Mail ACD event)

The other types of recordings (FTP and Overwrite Recording) do not support transcription.

A new action “Remove Skills” has been added to the group action “Change Parameter”. You can use this, for example, to remove required skills for a call if the caller has been in the queue for a certain amount of time, or if all agents have been tried but distribution was not possible due to the skill requirements of the current call.

It is now possible to setup tagging of call recordings made manually by agents (the agent starts / stops the recording) on an ACD Group or Client account basis.

See also this page for details: https://wiki.jtel.de/x/jAA8CQ

How to use:

• Setup recording tags, either in the ACD Group Parameters or the Client Parameters.
• For example: Recording.Tags.0 = Training, Recording.Tags.1 = Telephone Contract, Recording.Tags.2 = Nuisance Caller
• DO NOT USE COMMAS IN THE NAME OF THE TAG
• When calls are received, the button “Start Recording” is replaced with buttons associated with the tag names. In our example, three buttons would be displayed: Training, Telephone Contract, and Nuisance Caller
• When a recording is made, it is tagged with the tag associated with the button used to start the recording. Recordings started, for example, by pressing the Training button will be tagged “Training”.
• Multiple recordings can be made during a conversation (existing functionality), however, when the option “Always record agent” is selected, only one recording file is produced (stereo WAV file with the caller channel toggled on and off, as the agent starts / stops recording), and so only one tag can be set. If the option “Always record agent” is not set, each recording will receive an individual tag, according to the button used to start the recording.
• Recordings are saved to the portal as usual, and the tag is displayed in the list of recording shown via the menu option “Messaging … Call Recordings”.

Options

New options have been added to Client Master Data to govern how recordings are processed.

Upload call recordings with the following tags

• This setting is specified comma separated.
• In our example, if you specify Training, Telephone Contract as the value, only recordings tagged as “Training” and “Telephone Contract” would be uploaded to the FTP Server.
• If you specify nothing (no tags), then all call recordings will be uploaded if an FTP server is configured.

Call recordings are no longer visible in the Web UI after FTP upload

• If this option is specified, then call recordings which are set to be uploaded by FTP will not be available in the web UI via the menu item “Messaging … Call Recordings”

File Name Variables for Call Recordings

The following additional variables are available:

• $calldirection
  • OUTBOUND or INBOUND
• $connect_time, $connect_date
  • The time and date at which the call was connected with the agent.
  • The format is YYYYMMDD and HHMMSS respectively
  • Note: the variables $time and $date in contrast contains the time and date at which the call was initiated to the agent.
• $recording_tag
  • The name of the tag used when the recording was started (or empty, if no tag was used).
  • Caution: not all FTP servers will accept filenames with spaces gracefully. You will need to test on your FTP server to see if tags with spaces will be accepted, or use Tags without spaces.

Note on variable handling:

• Any variables which are added to the call process in the IVR will be available.
• So, if you declare a variable $customer_number in the call flow, you will also have access to this.

See also this page for details: https://wiki.jtel.de/x/cgEF

The group rules for creating the following types of events:

• Voice Mail
• Callback
• Fax

now support the setting of the event priority. Previously the priority for events created within ACD group rules was set to 50 by default.

Users who are members of a security group at the User level previously only had access to call recordings:

• for ACD groups of which they were an administrator
• and, if they were a member of one or more locations, only for those locations

Now users are additionally always given access to their own call recordings.

Note: if you do not wish users to have access to this feature, then the menu item “call recordings” should be removed from the corresponding security group.

For client administrators the behaviour is unchanged - they continue to have access to all call recordings in the client account.

A TTS file cache has been added to the system for the Azure TTS module.

The algorithm is as follows:

• Using the speech provider definitions (provider = Azure, speaker name, language) and the provided text or SSML, a SHA2 hash code is generated
• This defines the name of the file in the TTS cache
• After generation, the file is moved (a short time after usage) to the TTS cache.
• The cache directory contains 2 levels of sharding of the subdirectories. For example, if the cached filename is 5f64ccb4f704884442b41c3a4e1b9150b6b2f51ef43fefa503238add8602ef20.wav, then the file would be saved to the directory: \\acd-store\shared\Data\tts\cache\5\f

The next time TTS is performed:

• Using the speech provider definitions (provider = Azure, speaker name, language) and the provided text or SSML, a SHA2 hash code is generated
• This defines the name of the file in the TTS cache
• If this file exists, then the command is redirected to play the WAV file
• If the file does not exist, then the TTS is generated as usual (resulting in the file being stored in the cache)

Because the cache is stored on the data share, it will be covered by any backups made using the standard jtel backup procedure, and can be used by multiple telephony servers.

If you do not want to use caching, then set this setting to empty in the server.yml configuration file for blazar:

ttsCacheDirectory: //acd-store/shared/Data/tts/cache/

Note: caching will only occur, if the complete file is generated. If generation is interrupted during playback, then the file may not be complete, and will not be saved.

Sometimes when the MiniClient EXE is used, it is desired to communicate information to the jtel system on where a user has logged into - information such as the windows user name or the IP address of the workstation might be needed in the system to perform interactions with the workstation from a server or container in the jtel system.

The first use case for this is a particular CRM interaction where it is desired to have a screen popup on the local workstation when the agent receives an incoming call. The CRM system in question requires that a REST call be made to the local workstation of the agent to implement this functionality. This means, that the IP Address of the workstation must be saved in the Users jtel Portal data when the agent logs into the workstation.

See this page for an example of how to do this: https://wiki.jtel.de/x/dAQh

The new telephony server now supports SIP REFER. This is supported both in Agent Home / Mini Client and in the Workflow / IVR Object “Connect”.

The replace check box is now by default = false and also now works correctly.

An additional substitution variable is defined which can be used in any text templates provided via the confluence website:

${data:image/…} 

A regular expression is used to find text patterns that look like {{$

}}, where the content inside the curly braces begins with data:image/. Whenever such a match is found, it is replaced with an HTML <img> tag whose src attribute is set to whatever was inside the braces.

In other words:

1. Match: {{$ {data:image/...}

   }}

2. Replace with: <img src="data:image/...">

This can be used to embed images in the email, for example a graphic in the footer.

Note: the agent should not edit the data provided, this will undoubtedly break the image.

A new REST function provideCallData has been added. This can be used in scenarios where an external voice bot is connected and the voice bot wishes to provide information to the jtel system without disconnecting the call. https://wiki.jtel.de/x/T4VqAw

After call forwarding to a different group, recording settings are now determined by the destination group, ensuring forwarded calls do not inherit the recording option from the previous group.

When using Messaging ... Voice Mails, the filename for the download was always just "wav". This problem has been fixed.

A bug was fixed where the FTP call recording filename did not include the text "(unterdrueckt)" when the caller number was suppressed/anonymous. With this fix, you can now change the text that appears (previously "unterdrueckt") by modifying the value of the short text Caller.Restricted.

Conversations are now keyed by a composite identifier (chat_hash_sender) instead of only sender.
This change ensures that the same user can maintain separate conversation histories across different connectors without overwriting or mixing contexts.

The portal field size for the email subject field for jobs being sent by SMTP has been increased to 255 bytes. Now the subject will be truncated to 255 bytes if it is longer, instead of producing an error message.