LG CNS announced on the 28th that its artificial intelligence (AI) coding platform, "DevOn AI-Driven Development" (hereafter referred to as DevOn AIDD), maximizes productivity throughout the entire system development process. AIDD means optimizing the system development process through generative AI technology.
LG CNS has significantly upgraded the AI coding platform, which could only be used at the coding stage, to transform developer tasks across the entire system development process, from analysis and design to coding, testing, and quality diagnosis, into an "AI programmer."
During the system analysis stage, the AI coding platform summarizes the complex source code consisting of tens to hundreds of thousands of lines into natural language that humans can understand, greatly reducing the time to grasp and analyze the system structure. For example, when analyzing a bank's funds transfer system composed of thousands of lines of source code, it summarizes the entire service flow in natural language, including the following: checking the recipient's account number, verifying the entered transfer amount, checking transfer fees, verifying the withdrawal account balance, entering the password, and completing the transfer.
In the design stage, developers can easily design the system by adding or modifying sentences or words to the summarized content from the analysis stage. For instance, when adding an authentication method to the funds transfer system, they can simply write, "add voice authentication function among authentication methods" to implement the design. There is no need for developers to write out the system structure from scratch using design programs as before.
During the coding stage, AI maximizes the convenience of developers' coding tasks. Developers can enhance coding productivity by utilizing features such as: "code conversion," which converts past system codes to match the current version; "code generation," which automatically creates code by analyzing the context of existing code; "code recommendation," which suggests the next code to write during the coding process; and "code inspection," which detects and automatically corrects code that causes system errors, is vulnerable to security, or degrades performance.
The AI coding platform also increases efficiency in the process of testing the developed code. It automatically generates dozens to hundreds of test cases that developers previously had to create manually and precisely verifies errors by utilizing virtual data suitable for each case. In the quality diagnosis stage, it compares whether the completed code reflects the client's requests accurately and detects any missing items. It also assesses code quality regarding security vulnerabilities and bugs, suggesting improvement codes.
In particular, the platform comes equipped with a "coding agent" function that allows developers to input commands for the features they want to create, enabling the generative AI to automatically perform tasks from source code generation to testing and verification. When errors occur during the testing and verification processes, the coding agent automatically modifies the source code and repeats this process, resulting in high-quality coding outputs.
The AI coding platform can be broadly utilized, from small-scale projects to large system developments. LG CNS actively employs this technology in AX projects across various industries, including manufacturing and finance, innovating IT service execution systems and dramatically improving the quality of client systems. LG CNS aims to continuously upgrade the AI coding platform into an "agentic AI," where the AI autonomously plans and executes tasks across all stages of analysis, design, coding, testing, and quality diagnosis, merely upon the input of commands by developers.