Multiple reports say Amazon employees have been using an internal AI tool to raise company-wide “usage” or adoption metrics, rather than for essential work. According to the accounts, Amazon has been rolling out an in-house product referred to as MeshClaw, which lets staff create AI agents connected to workplace software and delegate tasks to them. The allegations, as described by outlets, claim that some employees use the tool to automate non-essential or unnecessary tasks in order to generate more activity on the company’s internal AI leaderboard or scoring system.
The reports characterize MeshClaw as enabling employees to assign jobs to AI agents that can complete work through integrations with internal systems. However, the coverage focuses on the disputed purpose of such activity—whether it reflects genuine operational need or is used to boost reported usage.
Both sources describe similar behavior and the same internal product. They do not provide evidence of broader company misconduct beyond employee reports, and they do not quantify how widespread the practice is.
