After Rippling blew millions on AI in months, it built an employee ROI tool
This title could be clearer and more informative.Try out Clickbait Shieldfor free (5 uses left this month).
Rippling discovered it was on track to spend 40% of its R&D headcount budget on AI tokens, with spend growing 80% month-over-month and one engineer alone burning $50,000/month. In response, it built AI Spend Console, a product that tracks individual and team AI spending, correlates it with actual productivity metrics (code output, pull requests), and routes prompts through an AI gateway to cheaper, task-appropriate models. After implementing the tool, Rippling cut token spend from 40% to 15% of headcount budget while maintaining nearly the same token volume — 600 billion tokens in July vs. 605 billion at peak — at 37% of the earlier cost, primarily by routing away from expensive frontier models toward alternatives like Z.ai's GLM 5.2. The product is included for Rippling HR subscribers and also available as a standalone.
Questions this post answers
How much did Rippling reduce its AI token spend after implementing cost controls?
Rippling cut its token spend from 40% of its R&D headcount budget down to about 15%. Token volume stayed nearly the same — 600 billion tokens in July versus a peak of 605 billion — but the cost of July's spend was only 37% of April's cost. The savings came primarily from routing prompts to cheaper, task-appropriate models instead of defaulting to expensive frontier models. Teams managing enterprise AI budgets track strategies like this on daily.dev.
What was the main cause of runaway AI token costs at Rippling?
Employees defaulted to using the most recent and most expensive frontier models for all tasks regardless of complexity. About 10–15% of employees drove 60% of total AI spend, and one engineer was spending $50,000 a month. Inference providers like Anthropic and OpenAI had no incentive to help control spend and provided limited usage insight, compounding the problem. Engineers and engineering managers navigating AI tooling decisions find real-world cost breakdowns like this on daily.dev.