The increasing use of Generative Artificial Intelligence (Gen-AI) in higher education offers a lot of opportunities to enhance collaborative writing. However, educators face challenges in understanding how students use these tools in group settings, which complicates their effective integration into the classroom. Moreover, little is known about the effects of specific Gen-AI applications. To fill this gap, the present study aims to identify particular Gen-AI applications in group writing tasks and examine how they relate to students’ group performance. The study focuses on analyzing how students use Gen-AI in group work, particularly when completing writing tasks. The sample includes 143 Korean students enrolled in an academic writing course, who were asked to collaboratively write a section of an academic paper. Based on students’ written self-reflections and identified Gen-AI applications in their texts, three categories of Gen-AI use were elicited: conceptual, language, and integrated. Conceptual applications involved tasks such as brainstorming, summarizing, paraphrasing, and other content-related activities. Language applications focused on grammar correction, vocabulary enhancement, and similar linguistic tasks. Integrated applications combined both conceptual and language uses. Conceptual and integrated applications were found to positively impact group performance, while language applications did not. These findings contribute to the broader discourse on Gen-AI use in education, with a particular focus on group-based applications and their effects.