July 10, 2026. Anthropic spent last week making Claude do more of your work. On July 9 it shipped a feature that shows you exactly how much. Claude Reflect is a built-in dashboard, in beta for Free, Pro and Max users with memory turned on, that tracks and visualizes how you use Claude: the topics you spend time on, your most active day and peak hour, the kinds of tasks you delegate, and observations about how you work. TechCrunch's read is blunt: it is also a retention play that makes you see how central Claude has become to your day, the way Gmail Meter once charted how much of your life lived in Gmail. Both readings are true, and the second one is exactly why the feature is useful to a business.
What Reflect actually shows
- A monthly recap. Topics, usage patterns, most active day, peak hour and observations about how you collaborate with Claude, under Settings, then Reflect.
- Reflection prompts. Occasional questions like what is one thing you want to keep doing yourself, even if Claude could do it faster.
- Time and focus controls. Optional quiet hours and break reminders, a nod to how sticky chat assistants have become.
- Usage coaching. Suggestions such as moving repeated context into Projects instead of re-explaining it, which also deepens the workflow lock-in TechCrunch flags.
- Privacy boundaries. Sensitive conversations appear only at a high level, anything connected to a health integration is excluded entirely, and Anthropic says insights data is not used for other purposes. A view of how much time you spend in Claude is planned for a later release.
What it means for operators
Most companies still cannot answer the basic question of where AI actually helps them, which makes every automation conversation start from guesswork. Strip out the wellness framing and Reflect is a free answer, an automation audit that runs itself. The measurement layer of the AI stack has now shipped at every level in barely a week: Anthropic's Cowork usage report showed which work people actually delegate, Enterprise admin analytics gave companies the org-wide view, and Reflect closes the loop for individuals. Use it that way: your recurring recap topics are your automation backlog, the tasks you repeat weekly are candidates for scheduled agents, and the context you keep re-explaining belongs in a project or a documented workflow. Run the recap across your team next month and you have a prioritized list of what to systematize. Turning that list into working automations is the part we do: see our AI automation services or talk to the agency team about an AI usage audit.
Frequently Asked Questions
Reflect is a dashboard inside Claude, under Settings, that tracks and visualizes your AI usage: topics, patterns, most active day, peak hour and observations about how you work with Claude. It launched July 9 in beta for Free, Pro and Max users and requires memory to be turned on.
Anthropic says sensitive conversations surface only at a high level, conversations connected to a health integration tool are excluded from insights entirely, and insights data is not used for other purposes. Reflect also adds optional quiet hours and break reminders under Time and focus settings.
Two reasons can be true at once. It helps users work more intentionally with AI, and, as TechCrunch notes, seeing your own dependence laid out in charts is a subtle retention play, much like Gmail Meter was for Google in 2012. The coaching tips, such as moving repeated context into Projects, also deepen workflow lock-in.
Treat the monthly recap as a free automation audit. Recurring topics are your automation backlog, repeated weekly tasks are candidates for scheduled agents, and constantly re-explained context should become a project or a documented workflow. Reviewed across a team, it produces a prioritized systematization list.