There is a problem that almost no one is talking about openly: businesses are spending serious money on AI tools, but most have no idea what they are actually getting for it.

Rippling found this out the hard way. By March 2026, the company’s CFO presented a sobering number to the executive team. Rippling was on track to spend 40 percent of its entire R&D headcount budget on AI tokens. And when they dug deeper, the picture got stranger. About 10 to 15 percent of employees were responsible for roughly 60 percent of total AI spend. One engineer alone was running up $50,000 per month.

The company did not fire that engineer. Instead, it built a tool to understand what was happening and whether it was worth it.

That tool is now available to other businesses. Rippling launched AI Spend Console on August 7, and it may be one of the more practically useful AI products released this year.

What AI Spend Console Actually Does

At its core, AI Spend Console connects three things that most businesses currently track separately, or not at all: what AI tools employees are using, how much they are spending, and whether that spending is producing measurable output.

The product sits on top of Rippling’s existing Data Cloud and Employee Graph, which already hold HR records, org charts, and identity data. That foundation matters because it means Rippling can tie token usage across tools like Claude, Cursor, and Codex to actual people, teams, and roles rather than just anonymous API calls.

CFOs and CTOs can drill down to see AI spend by department, team, or individual. They can also route requests to cheaper models where appropriate, putting guardrails on which tools different roles can access.

More importantly, the console connects AI spending to outputs in connected systems like GitHub and Salesforce. If an engineer is spending heavily on AI coding tools, the system can show whether their commit volume, PR throughput, or bug resolution rate reflects that. If a sales rep is using AI to prep for calls, you can see if their close rate moved.

That last part is genuinely new. Most AI spending tools stop at the cost side. Rippling is trying to close the loop on whether the spending is justified.

Why This Matters Now

The Rippling story is not unusual. It is actually representative of what is happening inside companies that adopted AI tools early.

In 2024 and 2025, many businesses gave broad access to AI tools without much governance. The logic made sense at the time: the cost of AI was low enough that the upside of any productivity gain outweighed the risk. Encourage experimentation, see what sticks.

By mid-2026, that calculus has changed. AI token costs have risen as models have gotten more powerful. Usage has expanded. And the original productivity assumptions have not always held up under scrutiny. Some teams are saving 10 hours a week with AI. Others are spending heavily and running slower than before because they are iterating endlessly on AI outputs instead of just doing the work.

Businesses need a way to know which camp each team is in. That is the problem Rippling is solving.

What This Means for Business

If you have deployed AI tools across your team, you almost certainly have uneven usage, concentrated costs, and unclear ROI. That is the normal state. It does not mean your AI adoption is failing. It means you are at the point where visibility becomes the next important step.

A few things are worth thinking about as AI spend tracking tools like this become standard:

Token costs are not the right unit of measurement on their own. The engineer spending $50K a month might be worth twice that if they are shipping features that would otherwise take a team of five. Or they might be generating a lot of impressive-looking AI output that no one is deploying. You will not know without connecting cost to outcomes.

Department averages hide what is actually happening. If your engineering team collectively spends $200K on AI tools and has solid output metrics, that looks fine. But if two people account for $180K of that and the rest of the team barely uses AI at all, you have a different problem on your hands. Visibility at the individual level changes what decisions you can make.

Governance does not mean restriction. The goal is not to cut AI spending. The goal is to understand it well enough to invest more in what is working and cut what is not. Rippling’s console includes routing logic that can steer requests to cheaper models for straightforward tasks, reserving expensive frontier models for work that genuinely needs them. That is a smart use of governance.

For most businesses, the first step is simply getting the data. Right now, AI spending is scattered across expense reports, corporate cards, and departmental budgets. Tools like Rippling’s give you a consolidated view that makes it possible to have a real conversation about AI investment strategy.

The Broader Shift

Rippling’s launch is a signal of where enterprise AI is heading. The early phase was about access. Get the tools in front of people. Enable experimentation.

The current phase is about accountability. Which AI investments are actually producing results, and which are just comfortable habits that have not been challenged?

That is a harder question. It requires better data. And it is the right question for any business that is serious about making AI work.

AI Spend Console is available to Rippling’s existing HR customers, with additional usage-based costs. It can also be purchased as a standalone product that integrates with other HR systems.