Anthropic launches tools to show businesses what AI spending buys

Anthropic has launched a new suite of reporting tools to give enterprise customers a detailed breakdown of what their AI spending on its models is buying, ahead of a planned IPO. The tools use internal analytics and employee surveys to show managers a team's work over a date range, its output, cost, and workflow failures. Anthropic says it won't use the data for model training by default and will offer recommendations for improving results, alongside new pooled group budgets for managing token spending. The move addresses a major concern among large customers about return on investment, as most Fortune 500 firms haven't seen major gains from AI. A Jellyfish report shows Claude Code spending grew 10x this year, now about 2% of engineering costs, with Anthropic maintaining its lead over OpenAI's Codex.