Anthropic launches Reflect feature to monitor Claude AI usage

Anthropic has released Reflect, a beta feature for its Claude AI assistant that provides users with detailed usage reports. The panel summarizes activity over 1, 3, 6 or 12‑month periods, showing frequent topics, peak usage times, conversation counts and the types of tasks performed, such as drafting documents or managing email.

Reflect also lets users set silent hours and break reminders, and it frames usage within Anthropic’s four‑dimension AI fluency model—delegation, description, discernment and diligence. Sensitive or health‑related conversations are excluded from the report, and no original files are accessed. The feature is available to Free, Pro and Max users who have memory enabled, and Anthropic plans to add total time‑spent views in the future.