Rippling launches AI Spend Console to tie token costs to output

Rippling has launched AI Spend Console, a product designed to track and control enterprise AI spending while linking usage to employee and team output. The HR software provider built the tool after it projected that AI tokens could consume 40% of its R&D headcount budget, amounting to millions of dollars.

Chief Product Officer Matt MacInnis said spending had been rising 80% month over month. At that rate, Rippling estimated token costs could approach 90% of the compensation paid to its R&D employees in the following year. CFO Adam Swiecicki presented the projection to the executive team in March, prompting an urgent internal review.

Usage data exposed concentrated spending

Rippling found that roughly 10–15% of employees accounted for about 60% of total AI spend. One engineer was using $50,000 worth of AI services each month. The company did not seek to end AI use; it sought to establish what it was receiving for the expenditure.

AI Spend Console maps spending by individual, team and role. Its dashboards combine measures such as prompts per day, spending and work output, including lines of code and pull requests. Rippling says the product can identify engineers with high AI costs whose peers regularly ask them to redo work during code review.

Routing becomes part of cost governance

Rippling first negotiated maximum spending caps with the tools it used, including Cursor, OpenAI and Anthropic. It also found employees were routinely selecting the newest and most expensive frontier models, regardless of the task. The resulting product includes an AI gateway that routes prompts to models intended to be more effective and cost-efficient.

Enterprises using a different gateway can still use AI Spend Console, MacInnis said, but its spend-governance features require Rippling’s gateway. The approach aligns with OpenAI enterprise credit analytics and spending controls, where enterprise controls can make usage data and spending limits more actionable across AI deployments.

Rippling reported that token spending fell from 40% to about 15% of its R&D headcount budget. Usage did not fall in parallel: the company reached a peak of 605 billion tokens in the month of the CFO’s warning and used 600 billion tokens in July. MacInnis said July’s token spending cost 37% of April’s because requests were being routed to more suitable models.

Productivity remains the deciding measure

The company also identified effective users and appointed them as AI captains to help colleagues. Engineering remains the main user group, while Rippling is working to apply AI to customer onboarding tasks such as mailing-data automation and data reconciliation. It expects the dashboard to measure results in terms of customers onboarded.

AI Spend Console is included for Rippling HR subscribers with additional usage-based AI charges, and it is also available as a standalone product that can integrate with another HR system of record. For businesses, the practical implication is to govern AI access through measurable productivity and task-appropriate model selection rather than token volume alone.