MentorChat: Introducing a Configurable Conversational Agent as a Tool for Adaptive Online Collaboration Support

Research in the field of computer-supported collaborative learning (CSCL) has systematically indicated that providing adaptive support to collaborating students is an efficient way of improving the outcomes of the collaborative learning activity. In this paper we introduce MentorChat, a web-based system that supports collaborative learning by utilizing a Conversational Agent. This agent, which can be easily configured through MentorChat authoring environment, aims to intelligently facilitate and trigger the discussion that takes place among partners. To achieve this, MentorChat models the domain through a series of key concepts and enables the teacher to define the form of agent intervention when a key concept is traced in students' dialogue. We present the system architecture and initial field evaluation data from two studies focusing on end-user acceptance of the system. We also discuss future system improvements and research agenda to explore the impact of MentorChat in collaborative learning settings.