Socar Modu Parking: Developing a Voice Parking Service in Two Days via AI-DLC Workshop

Socar's parking platform, Modu Parking, applied the AI-DLC (AI-Driven Development Life Cycle) methodology to develop a voice-based parking service MVP in just two days and launched its beta on September 2. By connecting Amazon Transcribe and Polly to the existing text-based agent, the team implemented voice input and response playback features, completing the system architecture using an orchestration server and an internal LLM router.

During development, the team switched to a turn-based conversation model due to technical limitations with voice interruption functionality. This required solving the challenge of precisely synchronizing the resumption point of the recognition stream between the app and the server. Additionally, contrary to the initial plan, the text chat feature was removed in favor of a voice-only UI to maximize usability while driving. The requirements documents and development rules generated through AI-DLC remain as project assets and continue to be utilized in future development.