Open Access

ARTICLE

Customer Service Support System: A Chatbot for University Reception

Muhammad Adeen Jamal¹, Bilal Khan^2,*, Sameed Ur Rehman¹, Wahab Khan¹

1 Department of Computer Science, City University of Science and Technology, Peshawar, 25120, Pakistan
2 Department of Computer Science, University of Engineering and Technology, Mardan, 23200, Pakistan

* Corresponding Author: Bilal Khan. Email:

Journal on Artificial Intelligence 2025, 7, 417-435. https://doi.org/10.32604/jai.2025.070762

Received 23 July 2025; Accepted 19 September 2025; Issue published 20 October 2025

Abstract

The development of artificial intelligence (AI) has sparked the invention of chatbots, which are intelligent conversational agents. These chatbots have the potential to completely transform how people interact while enhancing user experience. This study explores the building along with its execution of a chatbot for customer service support at a university reception using recurrent neural networks (RNNs). To increase user requests, the accuracy of the information, and overall satisfaction with the service, it evaluates machine learning models including RNN, XLNet, and Bidirectional Encoder Representations from Transformers (BERT). In this research project, data were gathered from university offices and students, documenting an array of daily questions that frequently arise at the main reception desk of the university. The recurrent neural network algorithm was trained using the gathered dataset, and it performed admirably. The model attained a low loss value of 0.0167 and an accuracy of 1.0000. The results presented demonstrate the efficiency with which the RNN model performed by precisely identifying and responding to the questions that were recorded. The thesis studies the pros and cons of RNN and measures how it performs when compared to the advanced XLNet and BERT algorithms. These systems’ efficiency can be gauged utilizing assessment metrics like accuracy, precision, and consistency of responses.

---

Keywords

---