Evaluate → Evaluate
AI-generatedSummary
Evaluates an AI agent's answer using the JaaS AI evaluation service, sending the question, answer, ground truth answer, context, evaluation criteria, evaluation type, and cohort to the API and returning the evaluation result.
Inputs
- chatInput — The question that was asked to the AI agent, taken from the incoming item JSON.
- output — The answer provided by the AI agent, taken from the incoming item JSON.
- context — Additional context for the evaluation, taken from the incoming item JSON.
- ground_truth_answer — The expected correct answer for comparison, taken from the incoming item JSON.
- cohort — Cohort identifier for grouping evaluations, taken from the incoming item JSON.
Output shape
object
Returns a single item with the fields below.
Examples
Example 1: Basic usage with only required fields
criteria1: Accuracy, type: Simple (S), question, answer, ground_truth_answer, context, cohort left as defaults or empty strings; no incoming data provided.
Example 2: Using incoming data to supply values
Leave all node parameters at defaults; previous node provides chatInput, output, context, ground_truth_answer, and cohort in its JSON.