In recent years, instant messaging applications have been very successful as one of the many technologies that people use to communicate and exchange information. Now, due to the significant progress in artificial intelligence (AI) and its sub-domains, such as Deep learning and Natural Language Processing (NLP), chatbots are becoming increasingly incorporated into these kinds of software applications. These chatbots are computer programs designed to communicate with humans through natural language. They are now widely employed in a range of areas, such as education and hospitality, to carry out certain tasks and offer ongoing assistance to users or deliver services to users by conversing with them in natural language. Recently, leading software development companies, such as Google, IBM, Amazon, and Microsoft, have expressed a strong interest in chatbot development. To respond to this fast-growing market, these companies have launched special products, such as Dialogflow, Watson Assistant, Amazon Lex, and Bot Framework, for the implementation of Chatbots. Although these frameworks and platforms exist, developing such software applications is a complex task requiring skills in various domains, such as software engineering, deep learning, machine learning, and NLP. Furthermore, one of the biggest challenges for these applications is to have a modeling language that supports a conversational scenario. Therefore, this paper aims to present a sketch of a new approach, based on the Unified Modeling Language (UML) to model a conversational scenario of chatbot. Then, it provides a case study of our approach by building a chatbot with the most suitable tool.