Customer Support is a person who serves prospective student questions appropriately and quickly. However, the conventional Customer Support requires high operating costs and has limited working hours. Moreover, in busy times, sometimes Customer Support is slow in responding to questions. In this study, a Chatbot system will be built to answer basic questions from prospective students, so it can improve the efficiency and effectiveness of Customer Support work at the Proklamasi 45 University. The more complex questions will still be directed to conventional Customer Support by the system. The Chatbot system in this study was built using the waterfall method, which consists of the requirements specification, design, implementation, testing, and implementation stages. The intent recognition will be performed using the StarSpace algorithm and Count Vectorizer by the Chatbot System. The frequency of occurrence of words in a sentence will be a reference to determine the purpose of a sentence. Testing is performed by Black box method by sending questions and evaluating the responses given by the system. The test results show that the Chatbot can provide answers with an accuracy of 97.75%. So it can be concluded that the Chatbot system can substitute the role of Customer Support. Accuracy can be improved by adding more data variations to the training data