Design coco-pilot : QFD as decision logic for conflict-aware human–LLM collaborative product conceptual design

Abstract

Intelligent product design significantly enhances daily convenience and improves work efficiency. Within intelligent product design, the conceptual phase serves as a critical foundational stage. Existing conceptual design methods typically require substantial time investment and human resources. Human–large language model (LLM) collaborative design offers a more efficient and cost-effective alternative by reducing reliance on human efforts. Meanwhile, it introduces potential design conflicts among multiple agents involved in the collaborative design process. To address these challenges, we propose an innovative framework called design Coco-pilot that integrates quality function deployment (QFD) into human–LLM collaborative conceptual design. The framework consists of three modules: interactive empathising through LLM-guided dialogue, automated and traceable construction of the QFD matrix, and QFD-guided multi-agent collaborative design. We apply the framework to develop the concept of the intelligent flight assistant. A multi-metric evaluation involving licensed pilots and aviation experts demonstrates the effectiveness of the proposed design. Our design Coco-pilot presents a novel approach that redefines user-centered design, facilitates intelligent product development, and explores new directions in human–LLM collaborative design.