Rippling 曾將 40% 的研發預算用於 AI，於是它造了個看門工具
ripples 削減了失控的 ai 成本（曾佔其 r&D 薪資預算的 40%），方法是推出一款追蹤員工 token 使用情況和輸出質量的工具。
rippling unveiled ai spend console, a tool designed to correlate ai token usage with employee productivity, following a discovery that its inference costs were growing 80% month-over-month. the company found it was spending as much on ai tokens as it paid 40% of its r&D staff, with one engineer spending $50,000 monthly. the tool attributes 'leaderboards' and features a gateway to route prompts to cost-effective models like glm 5.2 instead of expensive frontier models. by analyzing usage and enforcing spend caps, rippling successfully reduced its token expenditure without banning ai tools.