Description
API documentation is updated to reflect that the calibration property within the evaluation response only contains the id and selfUri fields.
Customers have reported encountering issues with missing evaluation calibration data from the API endpoint GET api/v2/quality/conversations/{conversationId}/evaluations/{evaluationId} as follows:
- The
expertEvaluator field is missing from the response of the conversations endpoint.
- All calibration scores (
averageScore, highScore, lowScore) return 0, despite them having variance.
Investigation revealed that when an Evaluation response includes a nested calibration object, the only valid fields within that calibration object are id and selfUri. All other fields are null or default primitive values (e.g., 0 for integers).
Change Category
Informational
Change Context
The current API documentation provides a misleading representation of the valid fields within the calibration object nested under the Evaluation response.
Change Impact
Customers will have access to updated API documentation accurately reflecting the valid fields (id and selfUri) of the calibration object nested under the Evaluation response.
Date of Change
14 Apr 2025
Impacted APIs
GET api/v2/quality/conversations/{conversationId}/evaluations/{evaluationId}
References
[SERVOPS-53008]
[QM-19315]