Rippling Launches AI Spend Console After Its Own AI Costs Ballooned
August 7, 2026
Rippling, a U.S. HR software company, launched AI Spend Console on August 6, 2026, giving companies a unified view of AI spending tied to business outcomes.
It breaks down AI costs by department, team, and role, and links them to work output.
It grew out of Rippling’s own wake-up call: the company’s token spending was on track to equal 40% of its R&D headcount budget.
How Rippling’s AI Spend Was Projected to Reach 40% of Its R&D Headcount Budget
Rippling’s official blog says the company opened AI tools to all employees early on, with no spending limits. Token spending grew 80% month over month, on a path to equal 40% of the R&D headcount budget — and nearly 90% the following year.
Chief Product Officer Matt MacInnis told TechCrunch that CFO Adam Swiecicki presented the numbers at a March 2026 executive meeting, leaving leadership incredulous.
An internal analysis found roughly 10-15% of employees drove about 60% of total AI spend; one engineer was spending $50,000 a month.
AI Spend Console: Cost Breakdowns Tied to Business Outcomes
AI Spend Console pulls usage data from vendors such as Cursor, OpenAI, and Anthropic and maps it to employee records. Companies can see which teams or roles drive their AI bill, link spending to performance ratings or pull request volume, and generate dashboards from natural-language questions.
Access controls for approved models and automatic request routing are planned for later release; the customer-facing Rippling AI Gateway is open for waitlist signups.
A 30-day free trial of AI Spend Console requires no Rippling subscription, per the blog. MacInnis told TechCrunch the console itself is included for HR subscribers, though AI usage-based costs apply. It is also available as a standalone product that integrates with other HR systems.
Spending Caps, Model Routing, and AI Captains
The results show up in the numbers. Rippling says spending once forecast at 40% of its R&D headcount budget is now held to 10-15%, savings of tens of millions of dollars a year. MacInnis told TechCrunch that July usage stayed near the peak — 600 billion tokens versus 605 billion — while July’s costs came to 37% of April’s.
The approach was straightforward: Rippling set monthly spending caps for Cursor, OpenAI, and Anthropic, then built an AI Gateway for its own internal use that routes all LLM traffic to the right model for each task. According to TechCrunch, CEO Parker Conrad said in July 2026 that Grok led its internal benchmarks while Z.ai’s GLM 5.2 delivered nearly identical performance to frontier models at 85% lower cost.
The company also named 20 “AI Captains” to support adoption beyond R&D. MacInnis added that, in his view, inference providers such as Anthropic and OpenAI have no incentive to help customers control spending.
Tomorrow Access previously covered a study estimating that only about 2,000 US engineers consistently deliver enterprise AI ROI.
What to Watch Next
Expansion beyond engineering is still a work in progress. According to MacInnis, Rippling is automating tasks for customer onboarding teams and plans to measure productivity by the number of customers onboarded.
He says wider access depends on linking token use in G&A (general and administrative) and customer-facing functions to productivity. Timing for the access controls and the customer-facing AI Gateway’s general availability is unannounced.
Sources
Rippling – From unchecked AI spend to complete control: How Rippling built AI Spend Console
Rippling – AI Spend Management and Optimization
TechCrunch – After Rippling blew millions on AI in months, it built an employee ROI tool
