Elicitron: LLM-Based Framework for Requirements Elicitation

This research article presents Elicitron, a novel framework that utilizes Large Language Models (LLMs) to enhance and automate the requirements elicitation process in product development. The paper outlines the challenges faced in traditional requirements elicitation methods, which often fail to capture the full spectrum of user needs, leading to products that do not meet expectations. Elicitron addresses these challenges by generating simulated users, or LLM agents, that engage in product experience scenarios to uncover a broader range of user needs, including latent ones. The framework is validated through three experiments, demonstrating its ability to generate diverse user needs, mimic empathic lead user interviews, and analyze interview data to identify and classify latent needs. The findings suggest that Elicitron can significantly improve the efficiency and effectiveness of the requirements elicitation process, ultimately contributing to better product design and innovation.