Anthropic launches Reflect, a dashboard for analyzing how you use Claude

On July 9, 2026, Anthropic launched Reflect in beta, a built-in Claude dashboard that shows patterns in how you use AI and helps users take a more deliberate approach to working with it.

What does the Reflect dashboard show?

The dashboard is available in Claude's settings on the web and in the desktop app. Once activated, it builds a personal report from your chat history over a period you choose: 1, 3, 6 or 12 months.

The report shows when you turn to Claude most often, which topics your work centers on, and what types of tasks you solve with AI. A total session time count is coming soon. Beyond the statistics, the dashboard periodically asks reflective questions, such as "What do you want to keep doing yourself, even if Claude could do it faster?", and offers to discuss your answer with the AI right away.

• Launch date: July 9, 2026, beta
• Availability: Free, Pro and Max users with memory enabled
• Analysis periods: 1, 3, 6 or 12 months
• Data covered: key topics, activity times, task types
• Exclusions: incognito chats and chats tied to health integrations

How Reflect helps build skills

Beyond analytics, the dashboard includes a learning framework, 4D AI Fluency, which helps users develop AI skills in four areas.

• Delegation: deliberately choosing which tasks to hand to the AI
• Description: describing a task precisely enough to get a useful result
• Discernment: critically assessing the quality and applicability of AI answers
• Diligence: personal responsibility for how and why the AI is used

The report shows how these qualities appear in your real work. For example, Reflect may note that you often rewrite email drafts in your own style, or that you formulate the strategy yourself first and only then delegate the details to Claude. The dashboard also gives practical tips: for instance, it suggests creating a Project so you don't have to explain the task's context from scratch in every new chat.

Quiet hours, breaks and privacy

In the dashboard you can set quiet hours, periods when Claude should not disturb you, and break reminders after a set amount of continuous work. Both tools are advisory: any notification can be dismissed with one click.

Anthropic paid particular attention to privacy. The dashboard does not pull source files from connected tools. If you asked Claude to summarize your email, only the summary reaches Reflect, not the emails themselves. Incognito chats and conversations tied to health integrations are fully excluded from the analysis. Data from Reflect is not used for any other purpose.

"In developing it, we worked with digital well-being experts from the AHA program at MIT

Media Lab, the Digital Wellness Lab at Boston Children's Hospital, and the Family Online Safety Institute," the Anthropic team says.

What this means

Anthropic has consistently pushed a responsible-AI narrative, and Reflect is another step in that direction. Rather than simply maximizing engagement, the company offers a self-assessment tool: do you see Claude as an assistant or a crutch? The feature is still in beta; support for Cowork chats will come later.

Frequently asked questions

How do I start using Reflect?

Open Settings in Claude on the web or in the desktop app and choose the "Reflect on your usage" option. If the report doesn't generate, check that the memory feature is enabled in your account settings.

Do private conversations and health data end up in Reflect?

No: incognito chats and conversations tied to health integrations are fully excluded from the analysis. Personal topics may appear only at a high level; specific messages are not included in the report.