AWS uses AI agent "Kiro" to revolutionize software development

The evolution of artificial intelligence technology is fundamentally changing the way software is developed. Deepak Singh, Head of Developer Experience at Amazon Web Services (AWS), emphasized at the recent “AWS re:Invent 2025” event: “Agent-based software development is no longer just a concept for hobbyist developers; it must scale to meet the rigor and maintainability required in enterprise environments.”

A recent hot topic, “ambient programming,” is a method where developers naturally generate code through conversations with AI agents. While this approach offers excellent accessibility and intuitiveness, it also has its limitations: the reasons behind code generation can become difficult to trace just weeks later, and it’s hard to share clear context within team development. To address these issues, AWS launched an intelligent development assistant called “Kiro,” which can transform conversational input into structured requirements documents and build automated workflows connecting testing and deployment.

Singh explained, “Just as veteran AWS engineers invest significant time defining problems and prioritizing code structure, Kiro can format user intent, which is then converted into executable work units after the design phase.” The resulting requirements documents not only enable the creation of simple unit tests but can also generate automated test cases based on functional requirements to verify whether the code operates as designed.

Additionally, Kiro can create user profiles tailored to specific roles (such as “AWS Ops Engineer,” “Frontend Developer,” etc.), automatically invoking the relevant toolsets and best practices. This allows security compliance and architectural standards to be embedded from the early stages of development, ensuring a certain degree of standardization even in large projects.

This type of agent-based development environment lays the foundation for maximizing productivity among senior engineers. Singh commented, “In the past six months, some senior developers have written more code than in the previous three years combined,” and “the integration of AI tools in development is truly unleashing their expressive capabilities.”

The emergence of AWS Kiro is transforming the paradigm of development models. It combines the free-form environment of ambient programming with the rigor of convention-driven structures, opening up new possibilities for building enterprise-grade software rapidly and reliably.