Learn how to evaluate AI voice agents using simulation, monitoring, STT benchmarks, semantic WER, latency, entity accuracy, and task completion.