Which of the following best describes the monitoring role in chatbot governance?

• A. Monitoring is optional.
• B. Monitoring helps detect drifting performance and triggers evaluation.
• C. Monitoring only affects uptime.
• D. Monitoring replaces the need for ownership.

Answer: (B)

Monitoring in chatbot governance is about spotting drift in performance and using those signals to trigger formal evaluation. Drift happens when how the bot behaves or responds changes over time—often due to new user phrasing, evolving data, or updates in content. By tracking metrics such as task success rate, intent recognition accuracy, user satisfaction, escalation to human agents, response latency, and safety incidents, governance teams can detect when the bot’s performance diverges from approved standards. When drift is detected, it prompts a governance evaluation—retraining the model, adjusting prompts, updating data quality controls, or revising policies, and sometimes enlisting human-in-the-loop intervention—to restore quality, safety, and compliance. This ongoing monitoring is essential for responsible AI and governance; it’s not optional, it covers more than uptime, and it doesn’t eliminate ownership—the designated owners remain responsible while monitoring provides the signals that trigger accountability actions.