Anthropic adds Claude smart reports to track workplace AI spend and activity

Anthropic has rolled out a suite of management features for its Claude Enterprise tier, led by a reporting tool called smart reports. The addition is designed to help organisations track what their employees are actually doing with generative AI and tally the associated computing expenses.

The feature works by having Claude analyse a sample of a team's recent conversation transcripts over a selected window of up to 28 days. The system then groups that activity into recognisable workstreams and deliverable categories, including software code, document drafts, data analysis, and technical summaries. Each category displays an estimated cost based on list prices, giving finance teams a direct look at the money spent on specific projects.

Beyond cost figures, the reports flag operational bottlenecks. Claude identifies instances where workflows stalled due to manual approval delays, missing system integrations, or prompts that failed to generate a usable outcome. The tool also provides a tally of interactions categorised as personal or off-topic, presenting those sessions strictly as aggregate numbers and costs without exposing underlying conversation details to managers.

Alongside the reporting dashboard, Anthropic added pooled team budgets and automated check-in surveys. Pooled budgets allow department leads to assign monthly usage caps to whole teams rather than managing individual seats, with alerts sent when spending crosses preset thresholds. The in-app survey module presents brief questionnaires to staff during natural pauses in their work, asking which tasks delivered the most value and which workflows could be adopted by other team members. Claude Enterprise accounts receive 10 free smart reports per month.

What this means for Australian businesses

Many Australian small and medium businesses struggle to measure the return on investment from workplace AI tools. While software subscriptions and token costs appear on credit card statements, business owners rarely have clear data on whether staff use those tools for productive work or casual exploration. Anthropic's new dashboard shifts AI management from guessing toward measurable accounting.

First, audit where your team spends its time in generative tools. If your firm uses Claude Enterprise, run a smart report across your main operational units. Check which deliverables consume the most computing credits and verify whether the finished work justified that expense.

Second, look closely at the recurring bottlenecks surfaced by the system. If multiple employees repeatedly abandon tasks because Claude lacks an integration with your customer database or accounting software, setting up those connections will save wasted effort.

Finally, update your internal workplace policies. While Claude masks personal names by default, the ability for management to inspect AI summaries and off-topic tallies introduces workplace privacy questions. Inform your staff clearly about how usage data is tracked and what safeguards protect their routine interactions.