Anthropic Launches AI Spend Reporting Tools for Enterprises

Anthropic introduced a new suite of reporting tools intended to help enterprise customers better understand the value of their AI spending, particularly for Claude Code and other agent-based coding products. The announcement comes amid strong financial signals from Anthropic, including an operating profit in the second quarter and 900% net revenue retention, indicating that existing customers significantly expanded their usage over the comparable prior period. The new tools give managers more detailed visibility into how teams use Anthropic’s products. Using internal analytics and employee surveys, the reporting suite can show what work was performed over a given period, what outputs were produced, how much it cost, and where workflows failed or underperformed. Anthropic said the data will not be used for model training by default, and the reports will also recommend ways to improve outcomes. The company is also offering pooled group budgets to help organizations manage token spending more effectively. The move reflects a broader shift in the enterprise AI market. As adoption of large language models and agentic coding tools accelerates, companies are no longer satisfied with simply purchasing access. They want clear evidence of return on investment. That concern is especially relevant as token costs rise across engineering teams. Recent data from Jellyfish, a developer productivity company, underscored the growing impact of Anthropic’s coding tools. Based on usage among 276,000 of its customers, Jellyfish found that Claude Code and other tools built on Anthropic models were responsible for nearly one-third of proposed code changes submitted for review. Spending on those tools has grown roughly 10 times since the beginning of the year and now represents about 2% of total engineering cost. Although that percentage remains modest, the rapid growth is significant enough to draw CFO attention and increase pressure on organizations to optimize AI deployment rather than simply scale it. Anthropic’s reporting tools are designed to address that need by turning raw AI usage into actionable business intelligence. For enterprise buyers, visibility into cost, productivity, and workflow bottlenecks can determine whether AI investments remain justified. For Anthropic, the tools reinforce its position as an enterprise-focused model provider by making its products easier to evaluate, manage, and standardize inside large organizations. The timing also matters for competition. OpenAI’s latest offerings, including Astra, have appeared to intensify rivalry in the frontier model market. However, Jellyfish said Anthropic’s adoption has continued to lead, and the expected disruption from rival tools has not yet translated into major customer switching. Part of Anthropic’s advantage appears to come from strong enterprise adoption and workflow integration. Once teams standardize development processes around a given model provider, switching costs rise because workflows, internal tooling, and agent behaviors become adapted to specific system quirks. Anthropic’s new reports may also serve as a lighter alternative to forward-deployed engineering, the embedded consulting model many AI labs use to help customers implement complex systems. By giving managers built-in analytics and optimization guidance, Anthropic is reducing the friction of enterprise AI adoption while strengthening customer retention.