Anthropic has introduced Reflect, a beta dashboard launched this Thursday (09), designed to allow users to monitor and analyze their interaction patterns with Claude. The purpose of this tool is paradoxical: it does not aim to increase the time spent with the artificial intelligence, but rather to help the user determine when usage should be reduced.
Dashboard features and comparison
This feature is accessible to free, Pro, and Max subscribers, provided the memory functionality is enabled. To use it, simply navigate to Claude's settings in the web interface or desktop application and select the Reflect tab. According to Engadget, Reflect's operation resembles features like Spotify Wrapped or Apple's Screen Time, as it summarizes what was done, when, and for how long, encouraging the user to question whether these activities align with their goals.
Usage analysis and user control
At the top of the dashboard, there is a textual summary of recent interactions, allowing the user to filter data from the last one, three, six, or twelve months. Below, data such as the day of highest activity, peak time, and the total number of conversations in the period are presented visually. Currently, the display of total time spent on Claude is unavailable, which is an intentional decision. Ryn Linthicum, Anthropic's Director of Wellness Policy, informed Engadget that the company chose not to maximize this data but plans to make it available in the future alongside usage management settings.
Break and limit settings
The dashboard offers the ability to configure quiet periods and set reminders to stop using the service after a certain interval. Both functionalities are presented as suggestions and can be ignored if the user is engaged in another task. Furthermore, Reflect proposes periodic questions to stimulate reflection, such as: 'What do you want to continue doing alone, even if Claude can do it faster?', whose answers can be debated directly with Claude itself.
Suggestions for efficient use
A practical section of Reflect consists of recommendations based on identified behaviors. For example, if the system notices that the user frequently needs to re-establish the same context when starting a new conversation, it will suggest creating a Project to group related interactions. These suggestions adhere to Anthropic's 4D AI Fluency Framework, developed in collaboration with researchers, which covers the dimensions of Delegation, Description, Discernment, and Diligence, all linked to the user's decision-making before, during, and after interacting with the AI.
Commitment to privacy
Linthicum reinforced her focus by telling Engadget: 'We were very intentional in building [the dashboard] with a focus on how we can improve people's usage, not in a way that encourages them to spend more time with it, but to help them be more efficient in achieving their goals—and, hopefully, leave Claude or preserve things they want to think about themselves.' In terms of security, Reflect excludes conversations conducted in anonymous mode, files linked by external tools, and conversations connected to health integrations. Anthropic guarantees that the information collected by the dashboard will not be used for any other purpose. The development benefited from consultation with specialists from the MIT Media Lab, the Boston Children's Hospital Digital Wellness Lab, and the Family Online Safety Institute. Availability for mobile devices and Cowork conversations is planned for future versions.