Generative Artificial Intelligence (GenAI), eminently through Large Language Models (LLMs), is swiftly transforming the landscape of many domains and in particular telecommunication networks by incorporating intelligent, data-driven automation and prediction. Beyond 5G (B5G) networks mainly aim to support massive connectivity, ultralow latency with high reliability and especially with the explosion of connected IoT devices and data-intensive applications several challenges persist. The main challenges due to massive number of connected devices including mobility-driven IoT applications like fleet management make the traffic dynamics increasingly complex, heterogeneous, and user-centric depending on the consumer requirements. To cater the need of the moment in B5G networks includes effective traffic prediction and load balancing which are critical factors for maintaining quality of service (QoS), optimizing resource utilization, and reducing operational overhead for the voluminous network load. In this chapter we present a comprehensive review of LLM-enabled service optimization for traffic forecasting and intelligent load balancing in B5G environments.