An in-home activity monitoring system for older adults has emerged to estimate the condition of an older adult and provide necessary assistance accordingly. Traditional approaches rely on detecting the daily activities of an elderly person and sending a notification to a caregiver when an anomalous event is detected. However, simply notifying an anomaly event is not sufficient for a caregiver to understand a source of abnormality and determine adequate assistance. This work presents an intelligent agent that can detect an anomalous activity pattern, but also support interactive dialogue responses to explain its detection and elicit missing information.