Anthropic has introduced a new set of reporting tools. According to TechCrunch, managers can use them to view a team's work over a chosen date range: what the team produced, what it cost, and where workflows broke down. The analysis draws on internal analytics and employee surveys. According to the company, this data is not used for model training by default. The reports are also meant to offer recommendations for improving results. Anthropic is additionally creating pooled group budgets so teams can manage token spending more effectively.
Background
Businesses are looking for this kind of information as they decide how to deploy AI tools. According to TechCrunch, usage among venture-backed firms is driving token spend, while most Fortune 500 companies are not yet seeing major gains from AI. The software company Jellyfish published a report based on agentic coding tool use by 276,000 of its customers. It found that nearly a third of all proposed code changes came from tools built on Anthropic models. Spending on those tools has grown tenfold since the start of the year and now equals about 2 percent of total engineering cost, according to Jellyfish.
Assessment
The development suggests a shift from simply spending on AI toward optimizing that spending. For German companies, this could mean that AI budget decisions are increasingly measured against demonstrable results, if such reports become widespread. It remains open how reliable the internal analyses and surveys are, since the data is collected by the vendor itself. Jellyfish's Nikolas Albarran also notes that the effort of switching makes it harder to move to another provider once a customer is established.
Sources
Editorially owned by Ideal Syka. Sources and method: Newsroom & method. Tips and corrections: ai@i6eal.de.
