AI 기반 의사결정 지원 서비스의 인지적 멀티모달 인터페이스 디자인 요소 연구: 통신 도메인 요금제 선택 맥락을 중심으로

초록

Background With the advancement of large language models (LLM), decision-support artificial intelligence has spread across industries. However, decision-support AI text-based interactions fail to reflect stepwise decision processes, increasing cognitive load. This study aims to design stepwise, multimodal user experience (UX) and to evaluate its effectiveness. Methods Based on a literature review, decision-making stages were defined, and a multimodal prototype was designed to match stage-specific cognitive demands using literature, service analysis, and in-depth interviews. Its effectiveness and user satisfaction were then evaluated through user testing. Results The multimodal interface showed higher satisfaction, modality suitability, and decision- support effectiveness. Visual structuring with auditory explanations improved comprehension, explicit presentation of recommendation rationale increased trust, and visual plus haptic feedback clarified task completion, supporting action transition. Conclusions Stage-based multimodal UX enhances decision-making support, addresses limitations of text-centered AI services, and provides guidance for cognitively aligned, integrated AI interfaces.

키워드