Rippling Was On Track to Spend 40% of Its R&D Payroll on AI Tokens, So It Built a Tool to Stop It

At a March executive meeting, Rippling's CFO revealed the company was on pace to burn tokens equivalent to 40% of its entire R&D headcount budget, with spend growing 80% month-over-month. One engineer alone was running up $50,000 a month. Ten to fifteen percent of employees were driving 60% of total AI spend. The culprit, per Chief Product Officer Matt MacInnis: employees defaulting to the newest, most expensive frontier models for every task, while Anthropic and OpenAI had "absolutely no incentives to help you control your spend."

Rippling's response was to build its own AI gateway and spend-tracking product, now called AI Spend Console, which routes prompts to cheaper models and maps individual spending against actual productivity signals like pull requests and code review outcomes. The results: token volume in July nearly matched the peak month, but cost 37% as much. The company's token spend dropped from 40% of headcount budget to roughly 15%. MacInnis summarized the routing logic plainly: "We're not letting the sales team do grammar updates using Fable."