Anthropic launches Claude intelligent reporting tool, allowing visibility into employee AI usage and costs
Anthropic launches the Claude intelligent reporting tool, available to users of the paid Claude Enterprise plan and can be enabled on the organization settings page. Administrators can select teams, a time range of up to 28 days, and the Claude products to include; the report generates charts showing usage, output content, and corresponding costs, and conversation details can be filtered by cost, product, date, and output type. Enterprises can generate up to 10 reports per month for free, with additional fees beyond that; Anthropic states that this feature should not be used to evaluate individual performance or for personnel decisions.
Why read this
AI score 52
The original text explains the report's selectable teams, time range, and product scope, as well as the entry point for adjusting quotas, making it easy to understand the specific form of enterprise-side AI usage management.
Topics
Body · AI translation
Anthropic Launches Claude Smart Report Tool to Monitor Employee AI Usage and Costs
English text is being prepared.
English text is being prepared.
English text is being prepared.
English text is being prepared.
Selectable products include Chat, Claude Code, Claude Cowork, or the "integrated Chat and Cowork interface." (The recently released Claude Duet AI agent is not yet supported for Claude Enterprise.) Smart reports extract data from the selected tools and summarize employee usage. The system reads team conversation records, categorizes the content into different workflows and output types, such as styled code and technical documentation, and also calculates the corresponding consumption costs.
Anthropic said: "If members cannot continue using the service due to quota limits, administrators with billing permissions can directly adjust the quota on the report page."
English text is being prepared.
Anthropic explained that, in addition to specifying the analysis focus of a smart report, managers can also customize query questions about usage and costs. These questions can be saved for repeated use, so they do not need to be re-entered each time. Smart reports also provide suggestions, such as recommending new rules that can be added to code repositories.
Each smart report generates interactive charts. Managers can click on the charts to view the details of individual conversations under a given workflow. Conversations can be sorted and filtered by cost, product used, date, and output type.
Smart reports also show the obstacles users encounter and how often they occur, while identifying what Anthropic calls inefficient usage scenarios, including conversations that produce no usable content and conversations flagged as personal use or off-topic from work.
Claude's new AI statistical tool is likely to make some employees uneasy. Multiple reports have mentioned companies using AI to decide layoffs, and a recent survey showed that one in four U.S. managers "often or always" uses AI to make such decisions. In July, a group of former Meta employees filed allegations claiming that Meta used insufficiently tested AI tools to assess employee productivity and track employees' AI token consumption in order to score employees as a basis for mass layoffs. Meta denied the allegations, saying that personnel decisions are made by humans, not by AI.
Anthropic appears to have noticed this kind of controversy and states that Claude smart reports are not intended for such uses.
Anthropic wrote: "Smart reports are for understanding tool adoption and planning Claude-related investment. They are not designed for, and should not be used for, evaluating individual performance or making personnel decisions."
Claude smart reports also impose restrictions on the questions that can be queried, not allowing users to query content related to race, gender identity, religion, political stance, immigration status, health conditions, union activity, harassment complaints, employees' private lives, and similar topics.