Anyone know why the Genesys Cloud Predictive Routing API is rejecting our batch upload of historical Zendesk satisfaction data? We are migrating a large Zendesk instance and trying to populate the Outcomes module with past CSAT scores to train the initial predictive models.
The request to /api/v2/predictrouting/outcomes/results/batch is returning a 400 Bad Request. The error message states “Invalid outcome type or metric configuration.” We have already created the Outcome Types in Genesys Cloud to match the Zendesk ticket fields, specifically mapping the 1-5 scale ratings. The JSON payload follows the structure outlined in the developer documentation for batch imports.
We are using the Python SDK version 2.0.1. The issue seems to stem from the timestamp format or perhaps the way we are referencing the Outcome Type IDs. Zendesk uses Unix timestamps while Genesys expects ISO 8601. We converted the dates, but the API still complains about the metric configuration. Is there a specific prerequisite in the Admin console for enabling historical data imports for Predictive Routing? Any guidance on the exact JSON structure required for the outcome results would be appreciated.