Amazon is encouraging its software developers to learn and apply AI in their daily work “as soon as possible”, saying engineers who fail to embrace the technology risk falling behind as software development continues to evolve.

In updated guidance for its software development engineers (SDEs), the company said developers already using AI should continue refining how they work with the technology because the pace of change remains high.

In the public memo to staff Amazon said it has deliberately avoided defining what good AI-powered engineering should look like, arguing that any guidance written today could become outdated within a year as the technology develops.

Instead, the company said it has revised its software engineering role guidelines to focus on what it believes will remain constant: the impact engineers have on customers, the business and their teams.

The guidance states that while the tools and techniques developers use will continue to change rapidly, the outcomes they deliver will remain the key measure of success.

The updated role guidance follows discussions between Amazon's distinguished and senior principal engineers about how AI is reshaping software engineering.

The company said AI is changing software development more quickly and fundamentally than previous technology shifts, including cloud computing, microservices and serverless computing. Despite this, it believes the core purpose of software engineers remains unchanged.

Amazon also cautioned against measuring developers solely by AI-related metrics such as token usage, AI tool adoption or lines of code. It said these metrics can encourage inefficient behaviour and fail to reflect the outcomes that matter most.

Instead, the company said engineering leaders should assess long-term results while using AI adoption metrics carefully to understand how teams are working. It added that teams making greater use of AI have generally delivered productivity improvements, while those adopting the technology more slowly risk falling behind.

The guidance also encourages developers to identify situations where AI does not improve their work. Amazon said engineers who can explain those limitations demonstrate the judgement the company expects.

Engineering leaders should also gain practical experience with AI tools themselves rather than relying on second-hand knowledge. Amazon said they should understand where AI adds value, where it falls short and help their teams adopt the technology effectively while continuing to focus on customer outcomes.

---