Anthropic's new Reflect feature nudges you to use Claude less

Anthropic has added a beta feature called Reflect that shows you how you actually use Claude. It sits in Settings and produces summaries of your recent activity: the topics you bring up, the kinds of tasks you hand over, and when you tend to reach for the model, across a window of one, three, six, or twelve months. Time-spent metrics are coming. The framing is unusual for a consumer AI product. Instead of pushing for more use, the tool asks questions like "What's one thing you want to keep doing yourself, even if Claude could do it faster?" and lets you set quiet hours or a reminder to take a break after a long session.

Anthropic ties the feature to what it calls a 4D AI fluency framework: deciding what to delegate, describing tasks well, judging the output, and staying responsible for it. The company says it built Reflect after user research found people want a clearer picture of how AI fits into daily life. The privacy limits are specific: incognito chats, files pulled in from connected tools, and health related conversations are left out entirely, and sensitive topics show up only as high level summaries. It is in beta now for Free, Pro, and Max users who have Memory turned on. The announcement has the rest.

Why it matters

A model provider giving you data to use its product less, and prompts to keep some skills for yourself, cuts against how most engagement features work. If you use Claude daily, Reflect is a low effort way to check whether the tasks you have delegated are the ones you actually meant to.