Amazon has pulled the plug on an internal leaderboard that ranked employees based on how much they used AI tools at work, and the official story doesn't match what workers are actually saying. The company claims the dashboard—called KiroRank—reached its goal of driving AI adoption across the organization. But multiple Amazon employees tell a different tale: the system was trivial to game, it rewarded waste over efficiency, and some workers deliberately flooded it with meaningless tasks just to climb the rankings.

How Workers Cracked the System

One employee described exactly how easy it was to cheat after receiving feedback that they weren't using AI enough during a performance review. Rather than integrate AI tools into their actual workflow, they started running automated prompts that generated endless busywork—tasks completely unrelated to their job. 'Honestly, iterating on that and maximizing the throughput was the most fun I've had at work,' the employee admitted. They didn't believe they were alone either, noting that internal discussions about wasteful spending suggest others caught on to the same workaround.

The Tokenmaxxing Problem

This incident highlights a growing phenomenon in tech: what insiders call 'tokenmaxxing.' Executives are increasingly measuring worker productivity by how much AI tooling employees consume, with some managers bragging about their AI spend outpacing actual headcount costs. It's a metric that rewards consumption over output—and when you tie visibility to those numbers, workers will inevitably find ways to maximize them regardless of whether the work matters. Amazon's internal PhoneTool award system (company-issued achievement badges) added another layer of gamification that encouraged employees to chase numbers rather than solve problems.

Amazon's Official Response

Amazon spokesperson told 404 Media that KiroRank 'was never intended to promote the use of AI for usage's sake' and described it as a beta dashboard that 'was not a formal or approved tool.' The company emphasized it doesn't mandate AI tool adoption across teams, though it does measure token utilization patterns for cost analysis. Internal reaction was mixed—some employees actually called for the leaderboard's return in comments on the shutdown announcement, suggesting genuine believers in the system alongside those who gamed it.

Key Takeaways

  • Amazon's KiroRank dashboard tracked employee AI usage and ranked workers accordingly
  • Workers easily cheated by running automated prompts with no work-related value
  • The incident reflects industry-wide 'tokenmaxxing' culture where AI spend equals perceived productivity
  • Amazon claims the tool accomplished its adoption goals, but insiders suggest waste concerns drove the shutdown

The Bottom Line

This is what happens when you gamify infrastructure costs—people optimize for the leaderboard, not the business. Amazon's 'success story' announcement reads like a quiet retreat dressed up as a victory lap. The real lesson here isn't that AI adoption worked; it's that surveillance systems built on vanity metrics will always get gamed by people who know how to push buttons.