Conversational agent as a support tool for learning C Language

A conversational agent is a software system that emulates communication with a human being through text or voice messages. The objective of this work was to evaluate the functionality and usability (effectiveness, efficiency, satisfaction) of a conversational agent called Codebot, , developed at the Faculty of Information Sciences, as a support tool to facilitate the learning of C programming language by students of the Computer Systems Engineering program, as well as their experience and hedonic (social) quality of the tool. The scope of the study was descriptive, with quantitative analysis. The data collection was crosssectional with convenience samples of different sizes. The System Usability Scale (SUS) was used, which was extended to evaluate three aspects: effectiveness and efficiency of the bot, as well as user satisfaction. The reliability of the instrument was 0.93 Cronbach's alpha. The behavior of the agent in the final functional test showed 70 % of successful responses and 30 % of unrecognized requests due to two main reasons: 1) the requested information was not in the agent's knowledge base, 2) poorly formulated or meaningless requests. Desing and implementation of the conversational agent Codebot obtained a robust database to improve learning of the C language and its function was satisfactory. However, is necessary to continue with tests that include a larger number of samples that delve deeper into the characteristics regarding the creation of the conversational agent.