Anthropic has shipped a meaningful set of management tools for companies running Claude at scale. The update is not about making Claude smarter; it is about giving administrators and finance teams a real picture of where AI money is going and the ability to act on it before costs run away.
The most useful new feature is a dashboard that shows spending broken down by team and by individual, placed directly alongside what each person actually produced. Admins can ask questions in plain language, such as which team doubled usage this month, and get charts back that can be shared in a board report. Usage data can also be pulled into financial tracking systems that companies already use, such as Datadog or CloudZero.
Anthropics new alerts notify administrators when spending reaches 75% and 90% of a set budget, giving them time to adjust before anyone gets blocked mid-task. Individual users can see their own spending trends and request a limit increase without leaving the platform. That removes a lot of friction and stops the situation where someone hits a cutoff at the worst possible moment.
The model controls let admins decide which version of Claude different teams can access. This matters because Claude has several tiers, and the most capable version costs considerably more per query than a lighter one. Letting routine tasks run on a less expensive model while reserving the premium version for complex work is a straightforward way to manage costs without cutting anyone off entirely.
All of this is arriving at a moment when the conversation inside most companies has shifted from "let us try AI" to "show us what we are getting for the money." Average monthly enterprise AI spending grew roughly 36% year over year between 2024 and 2025. Yet only 14% of finance chiefs surveyed recently said they have seen clear, measurable results from their AI spending. The pressure to prove value is real, and it is coming from boards and CFOs, not from IT.
Claude Enterprise pricing adds weight to the problem. The base fee is around $20 per person per month, but all actual usage is billed on top of that at rates that vary by model. Heavy users, particularly those running automated document processing or using Claude for coding work, can easily cost $100 to $250 per month or more. One analysis found that a 100-person team using Claude at serious volume could spend $66,000 per year once automated workflows are factored in, compared to $27,600 for moderate use. That is a 140% difference, and most finance teams only discover it after the fact.
The new tools do not solve the underlying complexity, but they make it visible in real time. For any organization that has approved an AI budget and is now waiting to justify it, the ability to show cost next to output, by team, by project, and by model, is the most direct path to that conversation. That is what Anthropic is selling with this update, and for the right audience, it is a credible offer.