Abstract:

In today’s competitive market, developing marketing campaigns is essential for success. The design thinking methodology has proven valuable in this domain but often requires significant time, resources, and a team of experts. Meanwhile, rapid advancements in Large Language Models have revealed their potential for both analytical and creative tasks, making LLM-based multi-agent systems a promising avenue for more complex problem solving. In this study, we propose an LLM-based multi-agent system that simulates the collaboration of human experts to create a marketing campaign guided by design thinking principles. We evaluate this approach with an experiment in which the system develops a campaign strategy for an eco-friendly beverage launch, highlighting key implementation aspects such as agent orchestration and prompt engineering. Our results show that the system can produce a comprehensive strategy by progressing through each stage of design thinking via collaborative ideation and research. Our research contributes to advancing the potential of agentic systems in business applications, demonstrating a capacity to deliver innovative and well-aligned marketing solutions.