When Rippling executives saw the numbers in March, the room went quiet. The company was on pace to spend on AI tokens an amount equal to 40% of its entire R&D headcount budget—millions of dollars—and the bill was climbing 80% month over month. That alarming discovery set off an urgent internal project that ultimately became a new product: AI Spend Console, an anti-"tokenmaxxing" tool the HR software provider unveiled this week to help companies track and contain their AI spending. Main Developments AI Spend Console maps how much individual employees, teams, and roles spend on AI tokens and whether that spending translates into genuine productivity gains—or just more "AI slop." The tool aims to show, for example, which engineers have high AI spend whose peers frequently ask them to redo work in code reviews, according to Rippling's blog post. Rippling's own analysis found that 10–15% of its employees drove about 60% of total AI spend, and one engineer was burning $50,000 a month. The company didn't want to stop AI usage, just rein it in dramatically. Read also: Why Poland's public websites are a hacker's playground Rippling's first move was negotiating maximum spending caps with each AI tool it used: Cursor, OpenAI, and Anthropic. That immediately exposed a common problem: employees defaulted to the newest, most expensive frontier models for everything, even simple tasks. Chief Product Officer Matt MacInnis was blunt about the incentive mismatch: "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." He noted they provide little usage insight and don't collaborate with one another. Rippling built its own AI gateway as part of the product, routing prompts to the most cost-effective model for each task. The console produces dashboards—once called leaderboards during the tokenmaxxing craze—that score attributes like prompts per day, work output (lines of code, pull requests), and spend. Results came quickly. Rippling dropped its token spend from 40% of its headcount budget to about 15%. In July, internal usage hit 600 billion tokens again—near the peak of 605 billion tokens in the month the CFO issued his warning—but the cost of July's token spend was just 37% of April's cost. "That's just because now we're routing to the more effective models," MacInnis said, joking that they're not letting the sales team do grammar updates using Fable. Background Rippling went all in on tokenmaxxing at the start of 2026—a widespread enterprise trend of aggressively adopting AI tools without much cost discipline. The company's March wake-up call came from CFO Adam Swiecicki, who presented the shocking numbers to the executive team. The launch ad for AI Spend Console even features Swiecicki sitting on a stool while employees dump wads of cash into a paper shredder—a nod to the moment that sparked the project. By mid-2026, enterprises had learned two key lessons, MacInnis said. First, they need multiple models from multiple AI labs at various price points, including cheaper open-weight options, possibly of Chinese origin. Rippling CEO Parker Conrad noted that in internal benchmarks, SpaceX's Grok was the all-around leader, but Z.ai's GLM 5.2 was 85% cheaper with nearly identical performance. (SpaceX now owns Cursor, which offers access to Grok and dozens of other models.) GLM 5.2 has become a favorite for coding tasks among tech companies, and Databricks has championed it too. Second, enterprises need an AI gateway to route prompts to the best, most cost-effective model for the task—exactly what Rippling built. Why It Matters The AI Spend Console highlights a broader shift: the tokenmaxxing era is giving way to cost-conscious AI governance. For Rippling, the tool is a critical step toward making AI accessible to all employees, not just engineers. MacInnis says technology alone isn't enough. Rippling identified employees who use AI effectively and made them "AI captains" to assist colleagues. But extending AI beyond engineering is a work in progress—so far, software engineers are the primary users. The company is piloting AI for customer onboarding teams to automate mailing data and data-reconciliation tasks, measuring productivity in terms of customers onboarded. "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 said. That means employee AI access may no longer be as ubiquitous as Slack or email—if productivity can't be measured, access might be restricted. What's Next AI Spend Console is included for Rippling's HR subscribers, with additional AI usage-based costs. It can also be purchased as a stand-alone product and integrated with another HR system of record, according to MacInnis. Enterprises that already use a