E-Prosiding Persidangan Serantau Manasik Haji 2025 (MANASIK2025)

MANASIK2025 Putrajaya, Malaysia, 7-8 Oktober 2025 651 of rituals, reminders for important steps, and reliable answers to common questions. This information is then used to design a system that not only meets technical standards but also respects religious authenticity and supports the spiritual journey of the pilgrims. The second phase is Chatbot Design and Development. In this phase, a knowledge base is prepared using trusted references such as the Muzakarah-Approved Hajj and Umrah Guidelines. This step is important to make sure the information given by the chatbot is correct and reliable. To enhance the system’s capability, a Generative AI model based on a Large Language Model (LLM) with integrated Natural Language Understanding (NLU) is incorporated. This allows the chatbot to interpret user queries in a more natural and context-aware manner. The model is designed to recognize the meaning and context of different questions, even if users ask them in different styles. The chatbot is also built to give clear and step-by-step answers, making it easy for pilgrims to follow and learn. For example, when a user asks, “What should I do after Tawaf?” the chatbot can guide them by explaining the next steps in simple terms. If someone asks, “Is it allowed to return to Madinah before completing Hajj?” the chatbot can provide an answer based on official rulings and guidelines. The system can also explain common mistakes during rituals and how to correct them. The overall design focuses on creating a smooth and friendly conversation that supports users in their Hajj and Umrah journey. By giving quick, accurate, and easy-to-understand answers, the chatbot can act as a personal assistant that helps pilgrims prepare before the journey and find guidance during the rituals. The third phase is Prototype Implementation, during which the AI chatbot is developed using Copilot Studio, Azure OpenAI, and a dedicated knowledge base. In this phase, the AI-integrated chatbot is designed to serve as a practical and reliable assistant to support pilgrims. The system is also planned to operate across both mobile devices and web platforms, ensuring accessibility through multiple channels and allowing users to interact with the chatbot in the way most convenient for them. The main goal of this phase is to make learning and practicing the rituals more convenient for pilgrims. The chatbot is designed with interactive features that help users step by step, for example showing the correct order of actions during Tawaf or Sa’i. It also includes a list of Frequently Asked Questions (FAQs) to answer common concerns quickly, such as what to do if a pilgrim forgets a certain step. In addition, scenario-based assistance is provided, which allows the chatbot to guide users through possible real-life situations, like being delayed in a crowd or missing a prayer time. Through these features, the prototype aims to support pilgrims in preparing themselves before the journey and in refreshing their knowledge during the rituals. By offering both convenience and reliability, the chatbot can serve as a practical tool that reduces confusion, builds confidence, and helps pilgrims perform their worship with greater peace of mind. Finally, the fourth phase is Evaluation, which involves usability testing with a sample group of prospective pilgrims, trainers, and subject matter experts. The goal is to make sure the chatbot gives accurate, clear, and Shariah-compliant answers, while also being easy and practical for pilgrims to use. The evaluation will be carried out in two parts. First, experts such as religious scholars and experienced trainers will test the chatbot by asking both common and complex questions about Hajj and Umrah rituals. These questions will include everyday queries like the correct steps of Tawaf, as well as more detailed cases, such as differences between schools of thought. The experts will then compare the chatbot’s responses with authentic Islamic references to check if the answers are accurate and in line with accepted rulings. Second, the evaluation will include prospective pilgrims, who will use the chatbot in a simulated training setting. Their role is to test how user- friendly and helpful the chatbot is in real scenarios. Pilgrims will be asked to judge the clarity of the explanations, the cultural appropriateness of the language, and the usefulness of the guidance in preparing for their journey. Feedback from both groups will be collected using structured forms, interviews, and rating scales. The chatbot will be assessed on key criteria such as accuracy of responses, user satisfaction, clarity, cultural appropriateness, and perceived usefulness. If weaknesses are identified, iterative refinements will be made to improve the system’s knowledge base, language style, and interface.