Currently, Anthropic is tackling a problem that grows as businesses adopt AI. Companies use AI to write software, support customers, and handle routine work. As adoption spreads, so do software fees, token charges and running costs.
Yet, a rising bill does not show whether AI improves company results. More prompts or faster output may look promising, but neither proves real gains. As a result, businesses need to know where AI helps, what it costs and whether gains justify more spending. Anthropic’s Claude Enterprise features aim to give managers a clearer view.
Anthropic Brings AI Activity and Spending Into View
On the 28th of October, the company released three features combining reports, staff feedback and budget limits. Anthropic’s Smart Reports review samples of team chats and group activity by task.
They show costs, work produced, and problems such as failed tool calls or missing app links. In addition, they flag repeated tasks teams could turn into shared skills and managers can ask questions and share findings widely.
However, chat data alone cannot explain a task’s full value. Anthropic’s in-app surveys ask employees which tasks help the business and save time, while Anthropic’s pooled group budgets let managers set monthly budgets while retaining personal caps.
Together, Anthropic’s tools give enterprises a broader view of adoption, but not proof of financial returns. The companies must compare those findings with measured business results.
Growing AI Investment Raises the Stakes
Consequently, this matters because companies are spending heavily on AI without regularly checking the gains. A September 2026 survey of 2,131 senior leaders across 20 countries found 89% reported measurable AI value. Yet, only 12% regularly compared gains with costs.
Moreover, software development highlights why stronger measurements are necessary. A study covered over 1,300 companies, 276,000 developers, and 100 million pull requests. It found that AI agents played a growing role in coding.
In September, AI agents opened or committed code in 27.7% of pull requests in an average company. This further highlights AI’s growing role and rising costs. But, more AI-generated code does not always mean faster delivery or lower software costs.
Turning Better Visibility Into Better Decisions
As a result, developers still need to review, test, and correct the output. Also, businesses must compare time saved with fees, infrastructure, training and human review.
Results will vary by task. AI may speed up coding but increase review time, or shorten support replies without resolving more cases. For enterprises, the question is no longer simply how widely employees use AI, it is whether those tools improve business outcomes enough to justify their growing costs.
Ultimately, Anthropic’s new features give companies a clearer starting point for answering that question. However, the financial case for enterprise AI will depend on measurable results, not adoption figures alone.
