fCrit: Towards an AI-Enabled Reflection-In-Action Interface for Furniture Design

Design insights are elusive. Capturing and transforming them requires a deliberate command of language through iterative sensemaking and reflection. Yet current GenAI use bypasses this intricate work, prioritizing automated content generation over supporting genuine creative sensemaking. We present fCrit, a conversational AI system that scaffolds formal analysis through adaptive levels of reflexivity. Through a preliminary pilot study with 13 designers, we observe three scaffolding features that may support aesthetic sensemaking while preserving creative agency: convergent reflection, divergent exploration, and mixed-initiative engagement. This formative study also surfaces an interface challenge: insights that emerge through dialogues are difficult to document and transform into interactive, tangible artifacts for users. To address this, we propose two interface features for future validation—manipulable insight environments and Feature Grab. We argue that AI systems for design learning should scaffold comprehension through interactive features before content production.