Chatbots and Socio-Emotional Cues in Corporate E-Learning: Evidence on Learning and Dropout

Abstract

Purpose: This study examined whether instructional chatbots and emojis improved learning outcomes and reduced dropout in asynchronous corporate training. It also explored participant performance profiles through agent-based simulation.
Methodology: The study followed three phases: chatbot development, implementation of a four-week private online course using a 2 × 2 factorial experiment (chatbot: yes/no × emojis: yes/no), and simulation in NetLogo. A total of 120 employees were evenly assigned to four experimental conditions.
Results: Chatbot use and emoji presence were associated with better learning outcomes and lower dropout. The chatbot-plus-emoji condition showed the highest survival rates (68%–76%) and the lowest dropout risk (24%–32%). Posttest scores indicated significant retention after 15 days (? = 6.61, p < 0.001). The simulation reproduced the empirical ranking of the conditions and identified a high-performing profile characterized by prior experience with virtual training, longer organizational tenure, familiarity with podcasts, and stable weekly study time.
Implications: The findings indicate that combining cognitive support through chatbots with socio-emotional signaling through emojis can strengthen persistence in corporate e-learning.
Originality: This study integrates factorial experimentation and agent-based simulation to explain how expressive digital interactions support learning continuity.