An Evaluation Metric is a measurable standard used to assess the accuracy, quality, or performance of an AI model. Different metrics evaluate speech recognition, language generation, latency, and conversational effectiveness.
Voice AI platforms monitor Evaluation Metrics to compare models, detect performance changes, and optimize customer experiences. Common metrics include Word Error Rate (WER), latency, task completion rate, and response accuracy.