Stuck on a recurring 500 Internal Server Error when initiating bulk export jobs for digital channel interactions that include bot transcripts. The issue arises specifically when the filter object includes metadata fields related to conversational AI interactions. Our legal team requires a complete chain of custody for all customer interactions, including those handled by virtual agents, so we cannot simply exclude these records from the export.
The API call in question is POST /api/v2/bulkexports/jobs. We are using the Python SDK version 2.3.1 and have verified that the JWT token has the necessary bulkexports:view and recordings:view permissions. The request body contains a standard filter for recording_type set to digital and a date range covering the last 30 days. When the filter explicitly includes transcript_text or attempts to pull bot-specific metadata, the server responds with a 500 error. The error message in the response body is generic, stating “Internal Server Error” without providing specific details about the malformed request or the server-side failure.
We have tested this with smaller date ranges and with filters that exclude bot interactions, and the jobs complete successfully. This suggests the issue is tied to the complexity or size of the bot transcript data. We are also integrating with an S3 bucket via presigned URLs, and the webhook configuration for job completion is valid. The problem persists across multiple environments, including our production and staging instances.
Has anyone encountered a similar issue with exporting bot transcript metadata? Are there known limitations or specific schema requirements for the filter object when dealing with digital channel bot interactions? Any guidance on how to structure the request to avoid this server error would be greatly appreciated. We need to ensure we are capturing all necessary data for our legal discovery requests without hitting these API roadblocks.