AI-Powered Conversational Agents in the Context of Learner Engagement in Engineering Education: A Systematic Review

Abstract

Online education is a critical part of today's education, allowing learners to learn on-the-go through digital platforms. But online learners often lack motivation, interaction, and ongoing support primarily because of the absence of human-like interaction in conventional online learning systems. To overcome these challenges, there is a growing trend in educational technologies to incorporate Artificial Intelligence (AI), such as conversational agents and virtual tutors, to offer interactive support and assistance. Conversational agents mimic human interaction and may serve as tutors, peers or assistants in online learning systems. Recent developments in Generative Artificial Intelligence (GenAI) and Large Language Model (LLM), such as GPT-based systems, have greatly improved their ability to interpret and respond to complex questions, and engage in multi-turn interactions. This paper is a systematic literature review of 54 papers from 2022 to 2025 on the elements and features of educational conversational agents. The review recognizes important inputs and outputs, recent AI approaches like Machine Learning (ML), Deep Learning (DL) and Large Language Model (LLM) approaches. It also discusses interaction approaches, Generative AI (GenAI) integration and metrics to assess learning performance, engagement, adaptability and technical aspects. The review suggests that the development of conversational educational agents should focus on long time personalization strategies with changing learner attributes and learning outcomes, integration of emotional intelligence, recommendation and ethical considerations