This action downloads transcripts from Genesys. It requires that the conversations have been downloaded first, either using the conversations() action, or using the conversations_from_datalake() action.
This action will store the transcripts of the conversations in the transcripts table. The transcripts are stored as a JSON array of objects, where each object represents a recording.
For voice recordings, the voice transcript is also downloaded and stored in the voice_transcript field.
The voice transcript is stored as a JSON array of objects, where each object represents a segment of the conversation.
Transcripts are also stored in the voice_transcript, chat_messages, email_messages, and message_messages tables, depending on the media type of the conversation.
You can define the start and end dates for the transcript downloads, or you can specify a list of conversation ids to download transcripts for.
The action will download transcripts for all matching conversations that span the start and end dates, or for the list of conversation ids provided.
A conversation spans the dates if any part of the conversation occurs between the start and end dates.
Genesys may not have transcripts available for all conversations, and Genesys does not always return a transcript, even if one is available. Genesys makes no guarantee about how long it will take to return a transcript,
and some transcripts may take longer than others. From time to time, Genesys may take many hours or even days to return a transcript, even if one is available.
If CX Vizz is unable to download a transcript, it will put it to the back of the queue, and re-attempt it up to 10 times. If after 10 attempts, the transcript could not be downloaded, it will be recorded in the
failed_transcript table. This will record the date/time the failure was initially noted, the conversation id, the number of attempts made to download the transcript, and the last time the download failed.
If you are downloading a large number of transcripts regularly, we recommend that you also run an additional scheduled job containing the failed_transcripts() action, which will retry any failed transcripts from the failed_transcript table.
This will ensure that any failed transcripts are re-attempted regularly, including transcripts failed due to Genesys not returning a transcript on previous runs.
Parameters
| Parameter | Description | |
|---|---|---|
| start_date | Optional | the start date of the range to download. Required unless retry_failures is true or conversation_ids has conversations. |
| end_date | Optional | the end date of the range to downloadRequired unless retry_failures is true or conversation_ids has conversations. |
| division_id | Optional | the division id to extract transcripts from |
| media_type | Optional | the media type to extract transcripts for. If not specified, all media types will be extracted. Should be one of Chat, Voice, Callback, Message or Email. |
| max_concurrent_downloads | Optional | the maximum number of concurrent downloads. If not provided, uses the configuration value MAX_CONCURRENT_DOWNLOADS or 20 if not set. |
| conversation_ids | Optional | a comma-seperated list of individual conversation ids to download transcripts for. If not provided, will download all matching conversations |
| download_attachments | Optional | if true, will download email attachments. Defaults to false. |
| queues | Optional | the list of queue-ids to restrict the download to. If not provided, downloads transcripts for all queues. |
Tables Affected
- transcripts
- failed_transcript
Variables Set
No variables set.
Permissions needed
- recording:recording:view or recording:recordingSegment:view
Examples
- conversations_from_datalake() transcripts( start_date=conversations_start_date, end_date=conversations_end_date)
- transcripts( retry_failed=true )