Designing with words: The role of prompt writing in AI-supported fashion visualization through Refabric

This thesis explores the integration of artificial intelligence into contemporary fashion design, focusing on the significance of prompt writing in image creation. As AI technologies become increasingly embedded in creative disciplines, the fashion industry is leveraging these tools for imaginative exploration and visual innovation beyond operational efficiency. Generative AI models, particularly those converting text prompts into fashion visuals, are redefining early-stage design processes by enabling rapid ideation and visual experimentation. Central to this process is the practice of prompt writing. The quality, clarity, and creativity of prompts significantly influence the accuracy and relevance of AI-generated outputs. Effective prompt engineering allows designers to control silhouette, pattern, fabric, texture, style and atmosphere, making it an essential skill in digital fashion workflows. By bridging conceptual thinking and visual realization, prompt writing enhances both creative direction and design efficiency. This study uses Refabric, a specialized AI-supported fashion design platform, as a case study. Refabric was selected for its focus on the needs of fashion designers, offering a tailored interface and models trained on fashion-specific data. Through critical analysis and practical experimentation, the research demonstrates how Refabric facilitates the generation of high-quality, design-relevant visuals and allows for iterative creative development through prompt manipulation. Ultimately, this thesis positions AI as a co-creative tool that expands the boundaries of fashion practice. It emphasizes the pivotal role of prompt writing in harnessing AI's potential and shaping the future of fashion design in the age of intelligent technologies.