Augmenting Divergent and Convergent Thinking in the Ideation Process: An LLM-Based Agent System

Generative Artificial Intelligence (GenAI) in general and Large Language Models (LLMs) in particular have recently gained considerable attention in innovation management as a means to augment the generation of innovative ideas. While this trend seems to grow at an astonishing pace, knowledge of how to leverage the transformative potential of LLMs for the generation of new ideas remains scarce in the scientific literature. This poses a major challenge for organizations striving to channel the inherent capabilities of LLMs for idea generation in a meaningful and efficacious manner. Against this backdrop, we design and instantiate an artifact that augments divergent and convergent thinking in the ideation process with the help of LLMs (i.e., LLM-based agent systems) following the design science research paradigm. Based on the insights from ten evaluation interviews with subject matter experts, we conclude that the integration of our artifact into existing ideation processes is useful and applicable.