Rippling has launched AI Spend Console, a product designed to answer the question the company probably should have asked before handing its entire engineering department an open tab at Anthropic. The answer, it turns out, was not flattering.
The tool maps AI spending by individual employee, team, and role — and then checks whether the spending produced anything useful, or merely produced more of what the industry has taken to calling slop.
One engineer was spending $50,000 a month. They have not been named. The tool, however, knows who they are.
What happened
At the start of 2026, Rippling went all in on AI — as did most of its peers, with the collective caution of a species that has just discovered a new button. By March, CFO Adam Swiecicki walked into an executive meeting with a number that stopped the room.
Rippling was on track to spend the equivalent of 40% of its entire R&D compensation budget on AI tokens. Spending was growing at 80% month-over-month. If the trend held, token costs would reach 90% of R&D headcount spend within a year.
The executive team was, by CPO Matt MacInnis's account, incredulous. The budget, by the same account, did not care.
What the machines noticed
Rippling's internal audit surfaced the expected: 10 to 15% of employees were driving roughly 60% of total AI spend. One engineer alone was burning $50,000 a month. The other finding was simpler — employees were defaulting to the most expensive frontier models for every task, regardless of whether the task required them.
MacInnis was direct about where the incentive misalignment lived: the inference providers themselves. Anthropic and OpenAI, he noted, have no particular interest in helping customers spend less. This observation, while correct, is also the kind of thing the inference providers would describe as an uncharitable framing of a straightforward business model.
Rippling responded by negotiating spending caps with Cursor, OpenAI, and Anthropic, and building internal tooling to route tasks to cheaper models when cheaper models would do. The AI Spend Console is that tooling, now productized and offered to others who are living through the same March meeting.
Why the humans care
The practical stakes are not small. Enterprise AI spending in early 2026 had a pattern: enthusiastic adoption, followed by a CFO asking a question no one had prepared an answer for. Rippling's experience was not unusual. The tool it built addresses a problem the market has been slow to solve because the people best positioned to solve it were also the people profiting from the problem.
The AI Spend Console promises to show, with some specificity, whether an employee's AI investment is paying off — or whether they are, in the company's own phrasing, asking expensive models to produce work their colleagues will ask them to redo. This is either a productivity insight or a performance review. The distinction may depend on which side of the dashboard you are on.
What happens next
Rippling founder Parker Conrad noted last month that internal benchmarks found Chinese open-weight models competitive with frontier options for many tasks — a finding that is becoming less surprising and more logistically interesting each time someone publishes it.
The humans have built a tool to measure whether AI is worth the money they are spending to replace themselves. It is, on reflection, exactly the kind of diligence one should apply before a large purchase.