Large U.S. companies are turning AI usage into a new internal performance indicator. As companies increase their investment in AI software and intelligent agents, management is demanding more explicit adoption data and pushing employees to integrate AI into their daily workflows.

Dashboard access to performance management

Companies like JPMorgan Chase, Meta, and KPMG have built internal dashboards to track AI usage by teams and individuals. Some systems are also made publicly available within the company, allowing employees to see how well their colleagues are using them.

For businesses, these tools serve both as a way to measure return on investment and as a means to identify employees who are slow to adopt them, thus influencing performance evaluations. Reports indicate that AI adoption is shifting from voluntary experimentation to more explicit job requirements.

Increasing call volume drives up costs

Once AI usage was quantified, some employees began trying to "boost their data." The article mentions that a practice called tokenmaxxing is emerging, which involves consuming more AI tokens to inflate an individual's ranking, even if these calls are not necessary for the job.

Such behavior is driving up costs for businesses and diminishing the relevance of dashboard data. Amazon previously had an employee-created AI token leaderboard, but shut it down in late May because the company was concerned that the mechanism would encourage employees to overuse AI for ranking purposes.

The monitoring scope has been expanded to include the entire work process.

The article points out that the ways companies monitor employees are also changing. Past workplace monitoring tools primarily focused on login time, mouse movements, screenshots, and online duration, especially after the widespread adoption of remote work.

Today, businesses are paying more attention to how employees write, code, communicate, make decisions, and complete tasks. One driving force behind this change is that AI agents are beginning to take on more tasks that require less human intervention.

For businesses, collecting this data not only helps determine who is using AI more effectively, but it can also be used to train the company's own AI systems. Meta stated in an internal memo in April that it would begin monitoring employees' mouse movements and keyboard input in order to use the data to train its AI system.

Bonuses and promotions are now linked

In addition to monitoring, some companies are also encouraging employees to increase their AI usage through promotions and bonuses. Accenture CEO Julie Sweet stated earlier this year that using AI has become a prerequisite for employee promotion.

KPMG has also launched an incentive program in its US consulting business, offering cash rewards to employees who generate new business ideas using AI. This practice demonstrates that companies' internal requirements for AI are shifting from simply "encouraging its use" to "incorporating it into performance evaluations."