AWS launches AI-DLC and the Kiro tool, targeting pain points in AI programming: “low efficiency” and “weak security”

Golden Finance reported that on the 16th, AWS Korea held a press conference at its Seoul headquarters and released an AI-driven development methodology (AI-DLC) and a Spec-based development tool called Kiro. The release aims to address pain points in current AI programming, such as “fast code generation, low delivery efficiency, and high security risks.”

AWS said that currently 94% of enterprises do not achieve expected results after introducing AI into software development. The reason is that existing AI tools mainly focus on the coding stage, and coding accounts for only about 20% of the entire development process. AI-DLC embeds AI across the full lifecycle—including requirements design, development, testing, and operations—so that, through continuous verification, human approvals, and end-to-end traceability, code, requirements documents, design documents, and operations manuals can be generated synchronously.

AWS said that several companies in South Korea have already achieved clear results: LG Electronics’ MS business unit has doubled its development efficiency; CJ Olive Young completed 5 project MVPs within 3 days; and the development cycle for large B2B platform projects was shortened from two weeks to one week.

At the same time, AWS launched the Kiro development tool, which can first generate requirements and design documents, and then proceed with code development. It also integrates functions such as automated testing, security checks, and connections to external systems.

AWS’s published test results show that under the same model, Kiro’s development cost can be reduced to one-sixth to one-tenth of that of comparable tools. (Jin 10)