Chatbot with Dialogflow for FAQ Services in Matana University Library

Abstract

Industry 4.0 and situation of Corona pandemic encourage the development of chatbots. The Matana university library in Indonesia is facing protocol implementation due to the pandemic, library staff vacancies and the need to improve 24x7 online services. Therefore, a solution is needed to resolve these kinds of problems. The success of chatbots in the business world and public services is the reason for the adoption of chatbots to overcome these problems. The first phase of chatbot development focuses on Frequently Asked Questions (FAQ). The development model for chatbot applications development uses a prototyping model by utilizing Dialogflow, namely the natural language understanding (NLU) platform. Agents, intents, entities, contexts, events, fulfilment, and integration need to be considered before designing a chatbot conversation. The test results are the chatbot is feasible to be accepted and operated. The variety of phrases entered into the Intent will improve accuracy. However, the accuracy of this chatbot will become lower if the language conveyed uses a lot of abbreviations or local / foreign languages. In further development, it is necessary to apply the system implementation and system maintenance phases, make chatbot responses not rigid like robots, add other library services such as new member registration, information on book availability, user satisfaction and so on.