Anthropic launches AI reports that track employee usage and spending
Claude's new smart reports allow managers to see how teams use AI, identifying 'off-topic' sessions and spending patterns.
The short version
- Reports analyze chat transcripts to group work into specific types of output like technical documents.
- Managers can filter sessions by cost, date, product used, and the specific type of output produced.
- The tool identifies 'inefficiencies,' including sessions where workers did not produce usable results.
- Anthropic restricts queries on sensitive topics like religion, health, and personal lives.
- Organizations get 10 free reports per month before incurring additional fees.
Why it matters
This tool gives employers a high-level view of how AI is integrated into daily workflows. It also introduces new ways for companies to monitor employee behavior and spending in real-time.
New oversight for enterprise teams
Anthropic recently announced the launch of smart reports for its Claude Enterprise plan. These tools are designed to give employers a clear view of how their teams are interacting with the AI. Managers can now see exactly why and how employees are using the technology on a daily basis.
The reports are available now and can be turned on within an organization's settings. They provide a visual overview of team activity through interactive graphs. These charts help leaders understand the volume of work produced and the total amount spent on the service.
As Mashable reported, managers can specify which products to include in these reports. They can choose to track data from chat, Claude Code, Claude Cowork, or a unified experience. The tool pulls this data from the chosen products to create a comprehensive overview of worker activity.
How the AI analyzes workstreams
The smart reports function by reading through team conversation transcripts. The system then groups these conversations into specific workstreams and types of output. This includes identifying the creation of technical documents or what Anthropic calls vibe code.
Managers can customize the analysis by selecting a date range of up to the last 28 days. They can also ask custom questions about usage and costs to get specific answers. These custom questions can be saved so that managers do not have to re-enter them every time they run a report.
The tool also provides proactive recommendations for the team. For example, it might suggest new rules that could be added to a code repository. This automated analysis helps managers identify where the AI is being most effective in the workflow.
Tracking costs and inefficiencies
One of the primary goals of the new tool is to monitor the financial side of AI adoption. The reports show how much money is being spent on specific tasks and products. If a spend limit is preventing a worker from completing a task, admins with billing permissions can raise the limit directly from the report.
Beyond just costs, the reports highlight what Anthropic calls inefficiencies. This includes identifying sessions where workers did not produce any usable output. The system also flags sessions that it classifies as being personal or off-topic.
Managers have the ability to click on interactive charts to see a deeper breakdown of these sessions. They can rank and filter these results by various factors. These factors include the date, the specific product used, the cost, and the type of output produced during the session.
Privacy and performance concerns
The introduction of these reports may cause some workers to feel apprehensive about surveillance. There have been reports of companies using AI to make decisions about layoffs or to rank worker productivity. A recent survey found that one in four U.S. managers use AI for this purpose often or all the time.
Anthropic has stated that these reports are not intended for evaluating individual performance. The company says the tool is meant to help managers understand adoption and plan their investments. Anthropic noted that the reports should not be used for making employment decisions.
To address privacy concerns, Anthropic has placed restrictions on the types of questions the AI will answer. Users are blocked from asking queries about sensitive topics. This includes questions regarding race, gender identity, religion, political affiliation, immigration status, health, and union activity.
The business context of the launch
This new feature arrives as Anthropic navigates its growth as an enterprise provider. In a recent IPO filing, the company noted that existing customers increased their spending by 900 percent since last year. This highlights the growing demand for corporate AI tools despite the company's reported losses.
The company reported a loss of $80 billion in the first half of 2026 alone. These reports suggest that while the company is still losing money, the adoption of its tools is accelerating. The smart reports are a direct response to the need for better data in these large corporate environments.
Organizations can generate up to 10 smart reports per month for free. After they reach that limit, the company will charge an additional fee. This structure allows smaller teams to experiment with the data while larger enterprises pay for deeper insights.
What to watch
- How companies balance productivity tracking with employee trust.
- Whether more AI providers release similar enterprise-level oversight tools.
Sources and links
- OfficialClaude Enterprise claude.com
- Officialisn't available on Claude Enterprise support.claude.com
- OfficialAnthropic wrote. support.claude.com
- OfficialAnnounced on Wednesday claude.com
- ReportingMashable: the article we sourced
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.
Listen to the AI Wire headlines
Listen to the AI WireListen to our daily AI shows every morning: on your commute, in the shower, wherever you are
Listen to the daily showsFirst reported by Mashable · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
