Today, we’re beta introducing a new way to reflect and improve how you use Claude. A common theme that emerged from interviews with users is a desire to better understand how exactly AI can be integrated into everyday life. How often should we use AI? How can we use it most effectively? When is AI appropriate for a task, and when is it best left to humans?
We built this feature to answer these types of questions. Easily track and visualize your Claude usage and determine if your time is meeting your goals. The Reflections Dashboard is located in Claude’s settings on the web or desktop app.
Check the pattern and adjust the shape
The reflection begins by summarizing how you used Claude, highlighting key topics, usage patterns, and types of tasks you often tackle.
You can look back at your Claude chat activity over the past 1, 3, 6, or 12 months.
The Reflection feature gives you a breakdown of when you used Claude the most and what you spent that time on. We’ll soon be adding a view of time spent using Claude.
Self-reflection also allows you to step back and examine the role Claude plays in your life. Questions like the following come up regularly: ”What would you continue to do on your own even if Claude could do it faster?” Please give me a chance to have a detailed discussion with Claude.
Within the dashboard, you can also set quiet times and schedules to remind you to take a break from using Claude after a certain amount of time. Both are reminders of your own preferences, so feel free to ignore them.
Build AI skills that support independent thinking
Using the 4D AI Fluency Framework, you can use your reflection to build new skills and refine how you work with Claude.
- Delegation: Set goals and decide whether and how to engage with AI
- Description: Effectively explain goals to drive useful AI behavior and output
- Identification: Accurately assess the usefulness of AI outputs and actions
- Diligence: Take responsibility for what you do with AI and how you do it.
The report provides an overview of Claude’s activity across aspects and provides examples of how you tend to collaborate with Claude. Examples include often reworking email drafts in your own voice, or delegating tasks only after deciding on a strategy yourself. Your reflections also offer practical suggestions, such as starting a project instead of recontextualizing the work in progress.
Privacy and sensitive topics
Reflections are never retrieved from secret chats, nor are the underlying files retrieved from connected tools. For example, if you ask Claude to summarize your inbox, the summary might be reflected, but you won’t see the source email. And conversations related to health integration tools are completely excluded from insights. The information and insights in your retrospective remain there. It will not be used for any other purpose. Learn more about our privacy practices.
We created this tool to reflect how Claude uses it. For some users, this may include sensitive or more personal conversations. To develop this tool and our approach to these topics, we collaborated with digital media and well-being experts at the MIT Media Lab’s Advancing Humans with AI (AHA) program, Boston Children’s Hospital’s Digital Wellness Lab, and the Family Online Safety Institute. Sensitive conversations can emerge as part of self-reflection, but only at a high level.
Start
The tool is currently available in beta for Free, Pro, and Max users with memory turned on. Claude on the web or desktop app[設定]and select the option that reflects your usage to generate the report. If you are unable to generate a report, it may be because memory is not turned on. Cowork conversation reflections will be available soon.
