Some employees at Amazon now seem to be using an internal AI platform called “MeshClaw” to automate extra workplace activity as pressure grows internally to demonstrate visible AI adoption to managers.
According to a report from the Financial Times, Amazon recently expanded the rollout of the in-house tool, which allows workers to create AI agents capable of interacting with workplace software and completing tasks on their behalf.
Inspired by “OpenClaw,” an open-source AI agent framework that gained widespread attention earlier this year for allowing users to run AI agents locally on personal devices, MeshClaw reportedly enables Amazon employees to automate tasks such as code deployments, email triaging, and interactions with workplace tools like Slack.
More than three dozen Amazon employees are said to have worked on the system internally, according to documents seen by the FT. One internal memo described the bot as something that “dreams overnight to consolidate what it learned, monitors your deployments while you’re in meetings and triages your email before you wake up.”
Amazon said the tool is already helping “thousands of Amazonians automate repetitive tasks each day” and described it as part of the company’s broader effort to encourage teams to experiment with AI technologies.
The Pressure Behind AI Adoption at Amazon
Even though the company has shared that these metrics are not part of performance reviews, employees say the visibility of usage data tells a different story. Some employees describe it as something managers are still paying attention to, which has made usage feel competitive rather than optional.
“There is just so much pressure to use these tools,” one Amazon employee told the FT. “Some people are just using MeshClaw to maximise their token usage.”
“Managers are looking at it,” another employee said, adding that, “When they track usage it creates perverse incentives and some people are very competitive about it.”
At Amazon, the pressure appears to have intensified after the company reportedly introduced targets requiring more than 80% of developers to use AI tools weekly, with some employees raising concerns about the security implications of allowing AI systems to take actions on behalf of workers inside company systems.
But according to employees familiar with the matter, some workers have started using the tool to automate unnecessary AI activity to increase their internal AI usage, particularly token consumption, which measures how much data employees process through AI systems.
Is AI Tokenmaxxing a Broader Silicon Valley Trend?
The situation at Amazon reflects a broader trend happening across Silicon Valley as major technology companies race to integrate generative AI deeper into day-to-day operations.
Companies including Meta, Microsoft, and Google are increasingly encouraging employees to adopt AI tools internally while also trying to demonstrate returns on massive AI infrastructure investments.
Amazon alone is expected to spend roughly $200 billion this year in capital expenditure, with much of that tied to AI and data centre infrastructure, according to the FT report.
The pressure to visibly engage with AI tools has likely created its own workplace culture with Meta, where employees are similarly engaged in what some internally call “tokenmaxxing,” where workers intentionally increase AI usage activity to improve their standing on internal dashboards.
While it does raise concerns among employees about how much autonomy tools like MeshClaw are being given inside company systems, some employees argue the unease goes beyond tracking and visibility and extends to how much control these tools actually have once they are running.
“The default security posture terrifies me,” one employee told the FT, adding that, “I’m not about to let it go off and just do its own thing.”
