LG CNS said on June 8 it launched DevOn Agentic AIND, an agentic artificial intelligence-based development platform that covers the full process of building and operating large IT systems.

LG CNS said “vibe coding” had limits for enterprise use, such as conflicts with existing systems or cases where partial changes affect the entire code.

LG CNS said it implemented a structure in which users enter requirements in natural language and specialised AI agents collaborate to carry out the entire development process end to end. These include an agent for analysing and designing customer requirements, a coding agent, and a testing and quality verification agent.

AIND performs customer-tailored development based on a “knowledge foundation.” The company said it applied to the development process a structured database of enterprise IT information that AI can understand, including development standards, security rules, source code and development outputs.

LG CNS said it also introduced a “spec-driven development” approach. It said the platform performs design, coding and verification based on pre-defined criteria to secure consistent quality regardless of the user and minimise hallucinations and errors.

AIND also supports “legacy modernisation” that automatically converts systems built in older languages such as COBOL to Java. It can shorten code analysis, conversion and verification work that used to take weeks or longer to minutes. It is currently applying the AIND-based “COBOL to Java” function to a next-generation project for a large South Korean financial company.

LG CNS said it jointly developed AIND with U.S. open-source AI coding company Cline. The companies plan to expand the application of AIND to global firms in security- and regulation-heavy industries such as finance, public services, manufacturing and defence, focusing on the United States, Japan and Southeast Asia.

Ahn Hyun-jung (안현정), an executive director in charge of application architecture at LG CNS, said, "Based on expert-level AI agents that understand enterprise systems, we will automate the building and operation of large IT systems and contribute to productivity innovation for corporate customers."