Full-scale Advancement of AI Coding Platform 'DevOn AIDD'
AI Performs Development Tasks Upon Command Input
'Coding Agent' Automatically Writes Code Based on Desired Functions
LG CNS is enhancing development productivity by introducing an artificial intelligence (AI) coding platform into every stage of its system development process.
On May 28, LG CNS announced that it will implement the AI coding platform 'DevOn AI-Driven Development (DevOn AIDD)' across all phases of system development. AIDD (AI-Driven Development) refers to the optimization of the system development process using generative AI technology.
Until now, LG CNS had only used its AI coding platform during the coding phase. The company has now advanced this into an 'AI Programmer,' which assists developers throughout the entire system development process?including analysis, design, coding, testing, and quality diagnosis.
First, in the system analysis phase, the AI coding platform summarizes millions of lines of complex source code into natural language that is easily understood by humans. This helps reduce the time required to analyze system structures. For example, when analyzing a bank’s account transfer system composed of thousands of lines of source code, the platform summarizes the entire service flow in natural language, including recipient account number verification, validation of the entered transfer amount, confirmation of transfer fees, checking the balance of the withdrawal account, password input, and transfer completion.
During the design phase, programs can also be designed using natural language. Developers can modify the content summarized in natural language during the analysis phase by adding sentences or words. For instance, if a developer wants to add a voice authentication feature to the account transfer system, simply writing 'add voice authentication function to personal authentication methods' prompts the AI coding platform to design the feature accordingly.
In the coding phase, the focus is on enhancing convenience for developers. They can utilize features such as: ▲'Code Conversion,' which transforms legacy system code into code compatible with the current version; ▲'Code Generation,' which automatically creates code by analyzing the context of existing code; ▲'Code Recommendation,' which suggests the next lines of code during the coding process; and ▲'Code Inspection,' which detects and automatically corrects code that could cause system errors, security vulnerabilities, or performance degradation.
The process of testing developed code is also automated using AI. The platform automatically generates hundreds of test cases that previously had to be created manually by developers, and it verifies for errors by applying virtual data tailored to each case. In the quality diagnosis phase, it checks whether the completed code reflects customer requirements and detects any missing items, notifying the developer. It also identifies potential security vulnerabilities or bugs in the written code and suggests improved versions.
Notably, the platform features a 'Coding Agent' that can automatically generate source code, test, and verify it simply by entering a command describing the desired functionality. If errors are detected during testing and verification, the Coding Agent automatically modifies the source code and repeats the process from code generation, enabling the delivery of high-quality coding results.
LG CNS plans to further advance its AI coding platform into an 'Agentic AI,' which autonomously plans and executes tasks at every stage of development when developers input commands. Ahn Hyunjung, Executive Director of Application Architecture at LG CNS, stated, "Through DevOn AIDD, we will revolutionize development productivity for our customers' systems and deliver a differentiated customer experience."
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
