Get Conversation transcript retrieves the messages from a Conversational Router conversation. Use it in an automation workflow that needs to process a past conversation, for example to pass its messages to another system.
This step is available in the automation workspace. You need the Conversational Router conversation ID for a conversation in the same account.
Set up the step
-
Add Get Conversation transcript from Get information in the step picker.
-
Under Conversation ID, enter the ID or insert the value supplied to the workflow.
-
Under Pipeline, select No pipeline for the original messages, or select a configured router pipeline to process them first.
-
If you need the pipeline's extra information about each message, turn on Include enrichment details.
-
Under Select output variable, choose where to store the transcript.
A pipeline can translate messages or remove personal details, depending on how it is configured. The enrichment option appears after you select a pipeline.
What to expect
The output contains the account ID, conversation ID, and message entries. Later steps read it from the output variable. The step retrieves message history; it does not start or reopen a conversation.
If a pipeline fails, the router can return the original messages. Selecting a pipeline that removes personal details does not guarantee that the returned text has been processed.
Enrichment details can contain personal information removed from the message text. Leave that option off unless the next steps need those details.
If retrieval fails, check the supplied conversation ID and account first. A successful request with no messages can produce an empty list.
Technical reference
If the output variable is transcript, its fields are transcript.accountId, transcript.conversationId, and transcript.entries. Message entries use the fields returned by the Conversational Router.
The step reads up to 20 pages of messages. Very long conversations can be truncated at that limit.