Amazon is tracking AI adoption among 2,100+ retail engineering teams, pushing some to triple code output while managing internal pushback

An internal Amazon document obtained by Business Insider reveals the company's retail division is monitoring AI tool adoption in granular detail across more than 2,100 engineering teams, tracking weekly deployments per engineer, monthly active users, net promoter scores, and a metric called “Value Deriving Event” that measures how frequently engineers generate outputs or provide feedback using AI tools. Amazon's goal is for 80% of its retail engineering teams to adopt AI-native practices — about 60% had done so as of February — with at least 25 teams expected to boost output tenfold this year, all tracked by Amazon's senior leadership team. The rollout has created internal friction, with engineers flagging concerns about top-down mandates, complex onboarding, overlapping tools, and a lack of clear success metrics, prompting Amazon to shift toward “collaborative AI practices” and automated reporting rather than requiring use of specific tools.