Rippling’s new AI Spend Console is aimed at a problem many companies are now confronting: AI tools are easy to adopt, but hard to manage once usage spreads across teams. The HR software provider built the product after its own AI token spending climbed fast enough to alarm its executive team.
The company did not decide to shut down AI use. Instead, it tried to answer a narrower and more practical question: which spending was helping people produce better work, and which spending was simply increasing the bill?
Why Rippling built an AI spending tool
At the start of the year, Rippling went deeply into what the source article calls tokenmaxxing. By March, Chief Product Officer Matt MacInnis said CFO Adam Swiecicki brought a number to the executive team that changed the conversation.
Rippling was on track to spend 40% of its R&D headcount budget on AI tokens. In plain terms, its token bill was approaching a level equal to 40% of compensation for employees in that unit. The source describes that as millions of dollars.
The trend was also moving quickly. Spending was growing by 80% month-over-month. If that pace had continued, Rippling expected the following year’s AI token spending could reach almost as much as 90% of what it spent on its R&D employees.
“We were incredulous,” MacInnis told TechCrunch.
That reaction led to what MacInnis described as an “urgent” internal project. Rippling wanted to understand not only where the money was going, but whether the spending was connected to real work output.
What AI Spend Console measures
AI Spend Console is designed to show how AI spending is distributed across a company. Rippling says the product maps spend by individual employees, teams and roles. It also tries to connect that usage with productivity, instead of treating all AI activity as automatically valuable.
One example from Rippling’s own description is especially direct. The company says the tool can show “which engineers have high AI spend whose peers frequently ask them to redo work in code reviews.” That matters because high token usage may look productive on a dashboard while still creating extra review work for others.
Rippling’s internal analysis found that “roughly 10–15% of our employees were driving about 60% of total AI spend. One engineer was spending $50,000 a month,” according to its blog post.
The product also creates dashboards that score attributes such as prompts per day, work output, lines of code, pull requests and spend. Earlier in the year, the source notes, these kinds of displays were known as leaderboards in the tokenmaxxing period.
How the company reduced token costs
Rippling’s first step was not to remove AI tools from employees. The company negotiated a max spending cap with each of the tools it used: Cursor, OpenAI and Anthropic.
It then found a straightforward source of waste. Employees were defaulting to the newest and most expensive frontier models for all kinds of tasks, even when those tasks may not have needed them.
“The truth is that the inference providers, like Anthropic and OpenAI, have absolutely no incentives to help you control your spend. They have every incentive for it to be a runaway expense, and that’s exactly what they do. They don’t provide you with great usage insight, and they don’t collaborate with one another,” MacInnis said.
Rippling also concluded that enterprises need multiple models from multiple AI labs at different price points. The source says that can include a frontier open weight option, perhaps of Chinese origin.
Founder and CEO Parker Conrad noted last month that Rippling’s own internal benchmarks found SpaceX’s Grok was the all-around leader, while “GLM 5.2 is 85% cheaper but [had] nearly identical performance” to the frontier models. The source also notes that Z.ai’s GLM 5.2 has become a favorite Chinese model for coding tasks among tech companies, and that Databricks has been championing it.
Why model routing became central
Rippling decided it needed an AI gateway that could route prompts to the most cost-effective model for each task. That gateway is now part of AI Spend Console.
MacInnis said companies that already use another gateway can still use AI Spend Console. However, if they want the features that govern spending, they would need to use Rippling’s gateway.
The results inside Rippling were significant. The company said it lowered token spend from 40% of its headcount budget to about 15%, while still keeping AI usage high.
MacInnis said internal usage peaked at 605 billion tokens in the month when the CFO issued his warning. In July, usage reached 600 billion tokens again, but “the cost of July’s token spend was 37% of the cost of April’s token spend,” he said.
“That’s just because now we’re routing to the more effective models,” he said, joking that “we’re not letting the sales team do grammar updates using Fable.”
What this means for employee AI access
Rippling’s approach suggests that broad access to AI tools may become more conditional inside companies. The issue is not whether employees use AI, but whether the organization can measure a link between token consumption and productivity.
The company also found that technology alone was not enough. It identified employees who were using AI effectively and made them “AI captains” to help the rest of the company.
So far, MacInnis said software engineers have been the primary users. Rippling is also working on use cases for customer onboarding teams, including automating some mailing data and data-reconciliation tasks. In that context, the dashboard would measure productivity by looking at onboarding more customers.
“We have to be able to link token consumption in G&A functions and in customer-facing functions back to productivity. If we can’t do that, all bets are off on any of this stuff being available to the broader employee base,” MacInnis says.
AI Spend Console is included for Rippling’s HR subscribers, with additional AI usage-based costs. MacInnis also said it can be bought as a stand-alone product and integrated with another HR system of record.