The telecommunications industry is increasingly reliant on human agents to handle complex customer support inquiries. With the rapid evolution of technologies like 5G and IoT, alongside a rapidly expanding customer base, there is a growing need for scalable, automated solutions that can provide accurate responses while reducing human intervention. This thesis aims to develop an AI-powered chatbot tailored to the telecommunications industry. The goal is to enhance customer support by delivering accurate, reliable, and up-to-date responses, using a sustainable and scalable architecture. The chatbot was built using a Retrieval-Augmented Generation (RAG) framework to dynamically integrate telecom-specific knowledge, avoiding the need for frequent retraining typical of fine-tuned Large Language Models (LLMs). The chatbot’s performance was rigorously evaluated using the RAGAS framework, demonstrating high accuracy, contextual relevance, and reliability. A comparative analysis with an open-source LLM confirmed the chatbot’s superior performance in handling telecom-specific queries. Real-world feedback via a survey from Ericsson’s BAM team further validated the system’s effectiveness, highlighting responsiveness while identifying areas for improvement, such as personalization.This research contributes to the telecommunications industry by offering a scalable, adaptable solution that enhances customer support while reducing the need for human agents. The approach can also be generalized to other industries requiring real-time, domain-specific responses, positioning the chatbot as a viable solution for reducing operational costs and improving customer satisfaction in complex, fast-evolving sectors such as telecommunications.