Performance Analysis of Generative Chatbot Integration on Message Communication

Integrating generative AI chatbots in communication platforms has become increasingly common; however, their effectiveness varies across different platforms. This study evaluates the performance of AI-based chatbots on WhatsApp, Telegram, and Discord, focusing on response time, response accuracy, and case-handling capability. This research adopts an experimental approach by testing chatbots on each platform using predefined academic-related queries, measuring response latency, BLEU-based accuracy, and the ability to handle document and link-sharing tasks. The results indicate that WhatsApp provides the fastest response time $(6.59$ seconds on average) but has limitations in delivering accurate responses and handling file transfers. Telegram achieves the highest accuracy (BLEU score of $\mathbf{0. 8 8}$) and a $\mathbf{1 0 0 \%}$ completion rate for case handling. Discord demonstrates moderate performance, excelling in interactive capabilities but lacking PDF file transfer support. In conclusion, Telegram proves to be the most effective platform for AI chatbot integration in academic services.