Abstract
The increasing complexity of airline operations and rising passenger expectations demand intelligent, scalable, and real-time customer support solutions. This research aims to develop a multi-modal AI-powered customer support agent for the aviation sector that enhances responsiveness and automation in customer interactions. The proposed system integrates Large Language Models (LLMs), function-calling APIs, and image-text processing to handle both textual and visual inputs, such as queries and boarding passes. A Gradio-based web interface enables seamless user interaction through text or image uploads, while backend Python functions manage core operations including flight status tracking, ticket rebooking, and baggage policy retrieval. The methodology emphasizes modular design and interoperability, allowing deployment with both proprietary (GPT) and open-source LLMs. Experimental evaluations demonstrate high accuracy, reliability, and improved user experience compared to traditional support systems. The study’s key contribution lies in presenting a practical and extensible framework for modernizing airline customer service through multi-modal, real-time AI assistance.