Anthropic Launches Reflect: A Dashboard That Shows Claude Users How They Actually Use AI

Anthropic has introduced a new dashboard inside Claude that lets users see exactly how they've been using the AI assistant — what topics dominate their chats, when they use it most, and how that usage lines up with their own stated goals.

The feature, called Reflect, launched in beta on July 9 and is available now through Settings on Claude's web and desktop apps. It works for Free, Pro, and Max subscribers, but only if they have Claude's Memory feature turned on, since Reflect draws its summaries from that stored history rather than from raw conversations.

Opening the dashboard generates a report covering the past one, three, six, or twelve months of activity. It breaks down the kinds of tasks a person tends to bring to Claude — coding, writing, research, planning, and similar categories — along with peak usage times and recurring topics. Anthropic has tied the reports to what it calls a 4D AI Fluency Framework, scoring interactions across four categories: delegation (whether and how someone chooses to hand off a task), description (how clearly they explain what they want), discernment (how they evaluate what Claude gives back), and diligence (how responsibly they act on it).

Beyond the summary, Reflect periodically poses questions meant to prompt self-reflection, such as asking users what's one thing they'd rather keep doing themselves even if Claude could do it faster. Users can also set quiet hours or schedule a nudge reminding them to take a break after a set amount of time; Anthropic describes both as dismissible reminders rather than restrictions.

On privacy, Anthropic says Reflect excludes incognito conversations entirely and leaves out anything tied to a connected health-integration tool. For other connected tools, the company says only Claude-generated summaries can surface in a report — if someone asked Claude to summarize their inbox, that summary might appear, but the underlying emails would not. Sensitive conversations aren't excluded outright, but Anthropic says they only show up at a high level rather than in detail. The company says the information generated stays within the Reflect feature and isn't repurposed elsewhere. Support for reviewing Cowork conversations is planned but not yet available.

Anthropic said it developed the feature's approach to sensitive material with input from the MIT Media Lab's Advancing Humans with AI program, the Digital Wellness Lab at Boston Children's Hospital, and the Family Online Safety Institute.

The launch has drawn comparisons to earlier usage-tracking tools from other tech companies, including Gmail's 2012 "Meter" feature, which showed users how central Gmail had become to their daily routines. Some coverage has noted that Reflect's suggestions tend to point users toward Claude's own features, such as starting a Project for ongoing work, rather than toward outside tools — a detail that has fed into broader questions about whether usage dashboards like this one are primarily aimed at helping users set healthier habits, deepening engagement with the product, or both at once.