Towards an Architecture for Educational Chatbots

Chatbots are gaining significant relevance due to their enormous potential in various sectors, particularly education. The COVID-19 pandemic led to suspending in-person classes for students and teachers, creating an opportunity for research in developing educational chatbots. However, the proposal of a widely accepted structure for chatbots poses a challenge due to the constant evolution of technologies and tools employed in their implementation and deployment. Moreover, generic chatbot architectures present limitations when instantiated in specific contexts. For instance, in the case of educational chatbots, it is necessary to have an architecture specifically tailored to the functionalities required in the academic domain. This article follows a formal approach to designing and implementing a software architecture for educational chatbots. We rely on an iterative method called ADD 3.0 to achieve this objective. Our proposal consists of a three-tier architecture to define the component distribution of chatbots and a six-layer architecture to specify their structure.