The operations team at Tailspin Toys notices that 40% of their Copilot Studio agent sessions end without the user's issue being resolved. The team wants to understand whether users are abandoning conversations early or whether the agent is failing to complete topics. Which metric helps distinguish between user abandonment and agent failure?
Select an answer to reveal the explanation.
Short Explanation and Infographic
Knowing whether users walked away or the agent hit a wall requires outcome-level data — resolved, escalated, or abandoned. Copilot Studio's session outcome breakdown separates these categories clearly. The correct answer is B.
Full explanation below image
Full Explanation
## Why B is Correct Copilot Studio Analytics tracks session outcomes in three categories: - Resolved: The agent completed the conversation and the user's issue was addressed - Escalated: The conversation was handed off to a human agent - Abandoned: The user ended the conversation without resolution (either they left early or the agent reached a dead-end)
Additionally, topic completion rates show whether individual topics are completing successfully — low topic completion on specific topics indicates agent failure, while high abandonment on topics that typically complete normally suggests early user drop-off.
## Why the Distractors Are Wrong A (Escalation rate): Escalation rate shows hand-offs to humans — it does not distinguish user abandonment from agent-side failures in topics that were never escalated.
C (CSAT score): Users who abandon early may not complete the satisfaction survey at all — CSAT requires a completed interaction to capture feedback. Abandonment and low CSAT are correlated but distinct signals.
D (Knowledge hit rate): Knowledge source hit rate indicates whether the agent found relevant content — it measures retrieval quality, not whether the overall session resolved the user's issue.
## Exam Tip Learn the session outcome categories: Resolved, Escalated, Abandoned. Also know that topic completion rate (within Analytics) measures whether individual topics run to their end node — useful for diagnosing agent-side failures.