Reflect With Claude: Guide to Anthropic's Usage Insights Dashboard

Reflect With Claude is a new beta feature from Anthropic that gives users a dashboard view of how they use Claude. It breaks down usage by topics, timeframes, frequency, and task types, helping people understand and shape their interactions with the AI.

Reflect also includes prompts, quiet hours, and gentle nudges for mindful use. Anthropic says privacy is maintained—incognito chats and certain integrations are left out of the dashboard’s analysis.

This guide explains what Reflect With Claude does, why it matters, how it compares to other tools, and best practices for getting value from your usage data.

What Is Reflect With Claude?

Reflect With Claude is a usage insights dashboard launched in beta by Anthropic on July 9, 2026, offering Claude users a detailed summary of their recent interactions with the AI.

The dashboard presents data for the last 1, 3, 6, or 12 months, showing key topics discussed, types of tasks completed, activity frequency, and temporal trends.

Reflect is designed to make AI use visible and intentional, helping users decide how and when to interact with Claude.

The feature also adds prompts for reflection—like which tasks you’d rather do yourself—and tools for managing engagement such as quiet hours and breaks.

Key Features and Data Tracking in Reflect

Reflect With Claude tracks usage patterns and summarizes user activity over selectable timeframes, allowing for granular or wide-angle review.

The dashboard highlights:

• A breakdown of recent Claude activity by topic (e.g. research, brainstorming, compliance, automation).

• Total number of chats and prompts issued in each period.

• Task types—such as drafting copy, paperwork automation, policy Q&A, or code generation.

• Visual charts showing how usage trends change with time (by week or month).

• Breakdowns by topic, task type, and frequency
• Selectable review windows: 1, 3, 6, 12 months
• Temporal trend visualization
• Prompts to encourage self-reflection
• Controls for quiet hours and break nudges

Privacy Controls and How Reflect Handles Data

Reflect With Claude is designed to respect user privacy by excluding private or sensitive content from its summaries.

Incognito chats (sessions launched in private mode), health integration conversations, and underlying uploaded files are omitted from analysis and display.

Reflections and prompts remain private to the user and are not visible to others or Anthropic support.

In operational client settings, we have seen privacy-excluding dashboards miss key workflow activities if staff use incognito for sensitive financial queries. Teams should clarify when to use each chat mode to balance privacy and oversight.

How Reflect Promotes Mindful AI Interaction

Reflect supports more intentional AI use by offering users prompts, behavioral nudges, and flexible engagement controls.

Reflection prompts ask users to consider questions like, 'What tasks would you prefer to do yourself?' or 'Are there areas where AI over-use could reduce skills?'

Users can set quiet hours (periods when Claude won’t send notifications or accept chat) and receive optional reminders to take breaks if activity becomes high-frequency.

For teams and SMBs, this can help prevent accidental over-reliance and nudge periodic review of AI workflows.

• Reflection prompts to support decision-making
• Quiet hours for uninterrupted focus
• Break notifications to prevent fatigue and over-use

Reflect With Claude vs Other AI Usage Insight Tools

Reflect With Claude differs from other AI usage dashboards by focusing on personal reflection as well as operational metrics, whereas other tools mainly address quantifying cost, activity, or governance.

Below, we compare Reflect With Claude, OpenAI's usage dashboards, and standalone AI analytic add-ons:

Examples, Tips, and Limitations Using Reflect

Small teams use Reflect With Claude to spot when AI is covering new domains (such as compliance queries) or if repetitive tasks could be delegated to automation instead of handled manually.

A typical use-case: a clinic owner sets a reflection nudge to check monthly that AI is not handling communications requiring clinical sign-off, using Reflect’s activity breakdown to identify drift.

One first-hand observation in small startup pilots: if teams do not regularly review their topic breakdown, inefficient prompt habits (like lengthy chain-of-thought for quick tasks) go unnoticed and compound over time.

Limitations include not analyzing incognito, health integration, or file detail—so patterns in excluded categories will not surface. Teams should discuss when privacy exclusion is appropriate and when a complete audit view is needed.