Vibe Coding: intention instead of implementation

Abstract Large Language Models (LLMs) and AI-assisted development tools are reshaping the way designers and UX professionals work. Under the term Vibe Coding , we describe a practice-oriented mode of working in which functional prototypes and software artifacts primarily emerge through iterative, natural language conversation with an LLM. Based on real-world projects, this article examines Vibe Coding from the perspective of two UX practitioners. Rather than representing a technological rupture, our observations suggest that Vibe Coding can be understood as a consistent extension of rapid prototyping. The key difference lies in the division of labor: instead of manual implementation, the emphasis shifts toward intent formulation, contextual steering, and continuous quality assessment. At the same time, diverse use cases reveal clear limitations. Without a defined UX process, user understanding, and explicit quality criteria, Vibe Coding quickly produces functional but generic output. Based on practical experience, we derive an Intent–Context–Quality model and discuss which competencies become newly important in Vibe Coding and which remain fundamentally central. This article presents a practice-based reflection grounded in real-world projects. It critically examines how Vibe Coding operates in professional UX contexts and outlines transferable insights for practitioners and researchers alike.