On July 9, 2026 Anthropic announced a beta feature for its Claude assistant called "Reflect," designed to help users review and refine how they use AI in daily life. The company said user interviews revealed widespread interest in understanding how often to rely on AI, which tasks are appropriate for automation, and where human judgment should remain primary.
What Reflect does
Reflect appears as a reflection dashboard in Claude's Settings on the web and desktop apps. It produces a summary of topics you engage with, your usage patterns, and the types of tasks you commonly handle with Claude. You can view activity for the past one, three, six, or twelve months.
The feature breaks down when you use Claude most and what you spent that time doing. Anthropic plans to add a view that shows the amount of time spent using Claude in a future update.
Reflect also prompts occasional reflective questions — for example, “What's one thing you want to keep doing yourself, even if Claude could do it faster?” — and offers the option to talk those through with Claude.
Within the dashboard you can set quiet hours or schedule a nudge to remind you to take a break after a certain amount of use. These are user-controlled reminders and can be dismissed.
Developing AI collaboration skills with the 4D AI Fluency Framework
Anthropic positions Reflect as a tool to help users build better workflows with Claude using the 4D AI Fluency Framework. The framework covers four dimensions:
- Delegation: setting goals and deciding whether and how to engage AI
- Description: effectively describing goals to prompt useful AI behavior and outputs
- Discernment: accurately assessing the usefulness of AI outputs and behaviors
- Diligence: taking responsibility for what we do with AI and how we do it
Reflect provides a summary of your activity across each dimension and gives examples of how you tend to collaborate with Claude — for instance, frequently reworking email drafts into your own voice or delegating tasks only after you establish strategy. It also offers practical tips, such as starting a Project to avoid repeatedly explaining ongoing context.
Privacy and handling sensitive topics
Reflect does not draw from incognito chats and does not pull in underlying files from connected tools. If you asked Claude to summarize your inbox, that summary may be reflected in the report, but the source emails themselves would not be exposed. Any conversations tied to health integration tools are excluded entirely from the insights.
Anthropic states that the information and insights in Reflect remain in the dashboard and are not used for other purposes. In building the feature, the company consulted digital media and wellbeing experts from the MIT Media Lab's Advancing Humans with AI (AHA) program, the Digital Wellness Lab at Boston Children’s Hospital, and the Family Online Safety Institute. Sensitive conversations can appear in Reflect, but only at a high level.
Availability
Reflect is available in beta as of July 9, 2026 to Free, Pro, and Max users who have Memory turned on. To generate a report, open Settings in Claude on the web or the desktop app and select the option to reflect on your usage. If you cannot generate a report, it may be because Memory is not enabled. Anthropic says reflecting on Cowork conversations will be added soon.
Anthropic describes Reflect as a tool to make users’ Claude usage more transparent and intentional, helping people align AI interactions with their personal and professional goals.
