Amazon has updated its guidelines for software engineers, shifting the focus of performance evaluations from AI usage metrics to tangible business value. Vice President and Distinguished Engineer Turk Turkovic stated that while AI is transforming development, the primary measure of success remains the results delivered to customers and the business.
Previously, Amazon tracked metrics such as AI tool usage frequency, token consumption, and lines of code, even creating leaderboards for teams. However, the company now warns that prioritizing these metrics can be counterproductive, leading to unnecessary AI usage that fails to improve software quality or customer experience.
Under the new guidelines, managers are required to use AI tools in real-world projects to better understand their capabilities and limitations. The company continues to emphasize AI as a core engineering practice, requiring teams to integrate it into their daily workflows to remain competitive.
Amazon will now focus on the top and bottom 10 percent of teams regarding AI adoption. Managers must investigate why the bottom 10 percent are not effectively utilizing AI and verify if the high usage in the top 10 percent actually accelerates delivery. Employees who cannot benefit from AI are encouraged to document their reasons, which the company views as a demonstration of professional judgment.
