Characterising Schön’s Reflective Practice within Text-to-Image GenAI-Design

Designers are turning to text-to-image Generative Artificial Intelligence platforms to support early-stage design work, yet little is known about how these technologies influence the designers’ reflective practices. This study investigates designer-GenAI interactions to describe how Schön’s Reflective Practice manifest in the designer’s use of text-to-image GenAI. From 25 protocol studies, we analysed 2,649 designer interactions with GenAI and the accompanying 9,704 GenAI image outputs. The analysis shows that prompts of high similarity were associated with reflection-in-action, while prompts of low similarity aligned with reflection-on-action. We also discovered branching explorations that suggest the presence of parallel reflections-in-actions, afforded by the speed of GenAI. These patterns reveal how GenAI outputs function as reflective material, making tacit design behaviours more visible with greater recording precision. This study advances our understanding of how designers adapt reflective practice to GenAI-design methods, and calls for further research on reflective methods with GenAI as it becomes more prevalent in design practice and education.

keywords: reflective practice; reflection-in-action; Generative Artificial Intelligence; Text-to-image GenAI

• with: Charlie Ranscombe, Nadia Anam, Nabila Afif, Awnili Shabnam
• year: October 2025 — ongoing