Token use is the new seat time.
At many companies, AI success is being tracked through usage dashboards: tokens consumed, seats activated, prompts per employee, weekly activity by team. Those numbers keep rising, and that growth is often presented as proof that a real transformation is underway.
Usage, adoption, effectiveness, and leverage are different things.β
β
A company can have high usage without changing how work gets done. It can integrate AI into daily workflows without improving outcomes. And it can improve speed on existing tasks without creating any meaningful strategic advantage.
Those distinctions matter because each one is harder to measure than the last. So many leadership teams stop at the easiest metric available: activity. Tokens are easy to count, easy to chart, and easy to present. But activity has never been a reliable proxy for value.
Most companies have made this mistake before. They have treated badge swipes as engagement and seat time as productivity. The number moved, the dashboard looked healthy, and the underlying work often did not improve.
AI raises the stakes because the distance between using the tool and using it well is massive and growing. A person generating huge volumes of low-value output can look far more productive on a dashboard than someone using AI selectively to remove friction, improve judgment, or eliminate work entirely. The metric rewards visible activity. It does not capture whether the work actually got better.
Output measures are harder to turn into a simple dashboard... and that is exactly why they matter.
β
AI adoption is only worthwhile if it improves speed, quality, or cost without meaningfully undermining the other two. Right now, most companies are celebrating gains in one dimension while ignoring losses elsewhere.
Measuring seat time as an input was a mistake. Treating token consumption as proof of value is the same mistake in a newer wrapper.
