A conversational agent (chatbot) is a software that communicates with people using natural language. Natural language processing and artificial intelligence rely heavily on conversation modeling. (AI).Indeed, creating a reliable chatbot system has been one of the most difficult challenges in AI since its inception. Although this system can do a variety of things, it must be able to recognize what users are saying in chat and reply to their questions. As a result, this study proposes a deep learning Arabic chatbot system built on a deep neural network model. The proposed method allows for text chats with others about common topics. The suggested system makes use of a Jordanian dialect derived from the website of Jordan’s Al-Zaytoonah University. The model is also created and tested in the deep learning frameworks Tensorflow and Keras. Furthermore, with no handcrafted rules, a dataset of manually produced pairs of conversations is used. This leads to the development of a chat bot system to assist users in counseling via the Al-Zaytoonah website. While the findings are promising and significant, the obtained results show that this system is capable of reproducing efficient and satisfying responses to users for the majority of queries. It is also demonstrated by the obtained results that an extreme burden is placed on the admissions department, including less online utilization of the website.