In this paper, we present ConvoGen: an innovative framework for generating synthetic conversational data using multi-agent systems. Our method leverages few-shot learning and introduces iterative sampling from a dynamically updated few-shot hub to create diverse and realistic conversational scenarios. The generated data has numerous applications, including training and evaluating conversational AI models, and augmenting existing datasets for tasks like conversational intent classification or conversation summarization. Our experiments demonstrate the effectiveness of this method in producing high-quality diverse synthetic conversational data, highlighting its potential to enhance the development and evaluation of conversational AI systems.
View on arXiv@article{gody2025_2503.17460,
title={ ConvoGen: Enhancing Conversational AI with Synthetic Data: A Multi-Agent Approach },
author={ Reem Gody and Mahmoud Goudy and Ahmed Y. Tawfik },
journal={arXiv preprint arXiv:2503.17460},
year={ 2025 }
}