[2Win5] Poster session 2

[2Win5-66] Developing Generative AI Prompts for Consumer Insights Based on Behavioral Observation Skills

〇Kazuya Nakamura¹, Ken Yasumatsu^2,3, Moeko Hamaguchi¹, Michiyo Endo¹, Yoko Maruoka¹ (1.OGIS-RI Co.,Ltd., 2.EVOLVE Inc., 3.Osaka Kyoiku University)

Keywords:GenerativeAI, Behavioral Observation, Prompt Development

With the growing adoption of design thinking and human-centered design, cultural anthropological approaches such as ethnography are increasingly practiced in various fields as methods to deeply understand people and foster innovation. In ethnography, "thick description," which captures not only human behavior but also the context surrounding it, is crucial for gaining comprehensive understanding. While generative AI-based research for consumer insights has been explored, conventional prompt engineering often falls short of generating practical, rich descriptions suitable for real-world research projects. This study aims to develop generative AI prompts capable of producing thick descriptions. Through four case studies utilizing generative AI, we found that interacting with LLMs using prompts informed by behavioral observation skills effectively generates rich descriptions that enhance consumer understanding. This paper presents the prompts developed for generative AI ethnography based on these practical applications.