A test case asking how to change a billing address passes Subagent, Action and Response Evaluation in Testing Center. A designer then scores the same transcript with the agent quality heuristics: the steps were correct but buried in six paragraphs, so the designer fails Conversational. Which statement best explains why the two evaluations disagree?
id q1fb2fe14cd