Amazon Urges Engineering Leaders to Address Slow AI Adoption

Amazon has directed its engineering leaders to investigate teams that are slow in adopting artificial intelligence, according to an internal memo. Published on October 27, the communication advised managers to focus particularly on the lowest percentile of AI adoption to understand the barriers involved.

The directive accompanied updated role guidelines for software development engineers, created by senior company engineers. While traditional technologies were previously excluded from these guidelines, senior staff concluded that artificial intelligence is altering the nature of work much faster than previous technological shifts.

The company acknowledged that current input metrics for tracking AI use, such as token counts and throughput numbers, are deeply flawed because they do not align with actual business value. However, leadership noted that heavily adopting teams are delivering increased output, while those failing to embrace the technology are falling behind.

At the same time, the memo assured engineers that it is acceptable if existing AI tools fail to improve their specific workflows, provided they document the reasons. Additionally, managers are expected to personally use AI on production work and identify both helpful and unhelpful tasks.