Evaluation → Evaluate
AI-generatedSummary
Evaluate an AI agent's response by sending a question, the agent's answer, and the expected correct answer (ground truth) to the RagMetrics API, receiving evaluation results and metadata in response.
Inputs
- Question (required) — The question that was asked to the AI agent.
- Ground Truth (required) — The expected correct answer for comparison.
- Answer (required) — The answer provided by the AI agent.
- Context — Optional context information to include in the evaluation.
- Conversation ID — An optional identifier for the conversation to which this evaluation belongs.
- Evaluation Group ID (required) — An identifier grouping this evaluation within a set of related evaluations.
- Type (required) — A single-character string identifying the evaluation type.
Output shape
A JSON object containing message, status, and optionally task_id, run_id, single_record_id, conversation_id, record_id, and results array.
The response includes a success status and may include detailed results if the evaluation succeeded. On failure, error information is provided if 'continueOnFail' is enabled; otherwise, the node throws an error.