Claude's new AI report card tool gives companies a closer look at costs — and flags 'inefficiencies'

Anthropic is giving company customers a new way to scrutinize how their workforce use its artificial intelligence chatbot Claude — and how much the AI activity is costing them.

The AI elephantine announced "smart reports" for Claude Enterprise users final week. The interactive diagram tool is designed to analyze a example of a team's AI meeting transcripts and collection activity into categories, including the types of tasks workforce use Claude for, the outputs it generates, and the cost.

The reports can additionally — and this is the part that is undoubtedly making teams sit up direct — emblem "inefficiencies" to employers, including sessions that produced "no activity value" or were "personal or off-topic," according to a Claude assistance webpage.

In effect, the astute reports tool offers companies armed alongside Claude Enterprise item of an AI study card. Each institution can run up to 10 reports a duration at no charge.

One caveat that Anthropic highlighted as "important" in its assistance documentation: "Smart reports assistance you comprehend acceptance and scheme your funding in Claude. They aren't designed and should not be used for evaluating idiosyncratic achievement or making occupation decisions."

To create a report, bosses can choose a team, a period range of up to the final 28 days, the Claude products to include, concentration areas for analysis, and any tradition questions they desire Claude to answer during analyzing the transcripts.

"Claude afterward says a example of transcripts in that scope, groups them into workstreams and types of outputs, attaches expend to all group, and writes up what it found," the Claude assistance leaf says.

According to Anthropic, sessions involving "restricted topics" — including individual conversations connected to race, ethnicity, health, lawful issues, union activity, and governmental or religious views — are "used lone for total inspection and don't contain summaries."

"Personal conversations additionally don't contain idiosyncratic sessions or names," the Claude assistance leaf says.

The initiate of astute reports comes as many companies are trying to get a grip on AI costs and measure any efficiency gains. Some are wrestling alongside how broadly to let workforce use AI tools.

Bloomberg reported Monday that Goldman Sachs lately constricted admission to Anthropic's Claude Code and Microsoft's GitHub coding tools to engineers lone following assessing efficiency and hazard concerns. The move reportedly rankled several Goldman traders who used the tools for their work.