CritiqueCrew: Orchestrating Multi-Perspective Conversational Design Critique

TL;DR

CritiqueCrew addresses cognitive load and cross-functional friction in UI design critique by embedding a multi-perspective, role-based critique workflow inside Figma. It combines three expert modules (UX, Product Vision, Engineering) with a Lead Coordinator to synthesize feedback and provide in-context remediation, transforming critique into actionable design changes. Two controlled studies (N=48) show that CritiqueCrew improves design quality, reduces cognitive load, and enhances trust and creativity relative to static baselines and unified expert models. The findings support a shift in AI’s role from a problem auditor to a solution co-creator, with practical implications for building future collaborative design tools that integrate explainable, cross-functional reasoning into designers’ workflows.

Abstract

UI designers face growing cognitive load and cross functional friction at the intersection of user needs, business goals, and engineering constraints. Existing automated tools often deliver static "problem lists", lacking actionable repair paths and disrupting creative flow. We introduce CritiqueCrew, a Figma tool that supports designers through conversational critique. CritiqueCrew generates multi-faceted insights by implementing a multi-perspective orchestration of distinct expert roles (UX, PM, Engineer). It translates abstract critiques into concrete actions via in context feedback and interactive remediation. Across two independent controlled studies (Total N=48), CritiqueCrew significantly improved both design quality and subjective experience compared to a traditional static checker. Furthermore, our results confirm that the structured orchestration of expert roles-rather than a unified model-is key to fostering trust and creativity support. Our work demonstrates how AI can shift from a "problem auditor" to a "solution co-creator" by integrating multi-perspective dialogue with interactive repair, offering design implications for future creative tools.

Figures (8)

• Figure 1: System architecture of CritiqueCrew
• Figure 3: Interaction workflow
• Figure 4: Comparison of design outcome quality between CritiqueCrew and the baseline tool in the realistic design task. CritiqueCrew led to broader issue coverage (left) and significantly higher solution effectiveness (right) than the baseline tool. Error bars represent standard errors. **$p < .01$, ***$p < .001$.
• Figure 5: Comparison of design outcome quality between the two CritiqueCrew modes: Unified Expert vs. Multi-Perspective. The Multi-Perspective mode yielded significantly higher issue coverage (left) and solution effectiveness (right), confirming the benefits of role-based critique decomposition. ***$p < .001$.
• Figure 6: Comparison of subjective experience metrics between CritiqueCrew and the baseline tool (SpecAI). (a) System Usability Scale (SUS, 0–100; Study 1 only). (b) NASA-TLX mental workload (0–100; lower is better). (c) Trust in AI (TAI, 1–5). Panel (a) is visually separated from panels (b) and (c) because SUS was collected only in Study 1, whereas NASA-TLX and TAI ratings pool responses from all study sessions in which participants used the baseline tool and CritiqueCrew. Across all metrics, CritiqueCrew was perceived as more usable, imposed lower mental workload, and was trusted more than the baseline tool. ***$p$ < .001.