Anthropic introduces 'Smart Reports' for Claude Enterprise, offering insights into AI usage, costs, and inefficiencies as it prepares for a $200B IPO.
New feature for AI tool
Anthropic, a U.S. company focused on artificial intelligence, has introduced a feature called "Smart Reports" for its Claude Enterprise product. This feature helps organizations using Claude Enterprise to examine AI-related activities, evaluate task costs, and identify recurring issues in their workflows.Background on Anthropic
Anthropic was founded in January 2021 by former members of OpenAI, including Daniela Amodei and Dario Amodei, who are the president and CEO, respectively. The company is preparing for an initial public offering (IPO) in November 2026 and is currently private.Functionality of Smart Reports
The "Smart Reports" feature allows authorized users to select a team, choose a reporting period up to 28 days, and specify which Claude products to analyze. Claude then analyzes a sample of conversation transcriptions, grouping sessions by workflow and result categories. These reports can cover activities such as documentation production, code writing, and analysis, along with associated costs.Usage and limitations
Managers can use the reports to review work, track costs, and identify recurring problems like tool failures or repeated revisions. The system can suggest workflow improvements and allow administrators to adjust spending limits. However, individual performance evaluations or employment decisions are not intended uses of the reports.Privacy and protection measures
Conversations about restricted topics are included in aggregated reports but not individually summarized. Individual attribution is disabled by default, and if enabled, the names and session identifiers remain masked until a user reveals them. The system also limits questions about sensitive topics such as health, religion, and employee privacy, and these protections cannot be disabled.Purpose and considerations
According to Anthropic, the Smart Reports help organizations understand AI adoption and plan their investments in Claude. The reports show how teams use AI, the associated costs, and which sessions are considered unproductive. Transparency and careful interpretation are emphasized, as unproductive tasks may reflect technical issues rather than wasted time.Additional features
Anthropic has also introduced Surveys and Shared Budgets for Claude Enterprise. Surveys collect user feedback on how Claude helps with tasks, while Shared Budgets allow administrators to set group budgets per team. These features are now available to all Claude Enterprise users.IPO and market skepticism
Anthropic plans to go public in November 2026, aiming for a valuation of $200 billion. However, an independent research firm called "New Constructs" has criticized the valuation, calling it the "most ridiculous IPO of 2026" and estimating the company's worth at only $150 billion. New Constructs also questioned the viability of Anthropic's business model, noting that it would need to record profits far exceeding those of major competitors like Nvidia.