5G networks introduce complex challenges in mobility management and energy consumption, including more frequent handovers as users move between cells, network congestion, service disruptions, increased power consumption due to higher path loss, and the need for beamforming. This work performs an experimental validation on optimal mobility management and energy consumption efficiency in 5G networks fuelled by Artificial Intelligence (AI) models. To achieve this, two real-world experiments in a 5G testbed were conducted, assessing the: (i) dynamics of handovers (HOs) between gNodeBs (gNBs) within a single operator network; and (ii) energy consumption characteristics of 5G base stations under various traffic conditions. To advance location-aware and energy-saving intelligence in 5G networks adaptive AI-fuelled policy-enforcement mechanisms was leveraged. The results of this work demonstrate significant insights for network operators which aim to reconcile performance network demands with energy consumption limitations, and develop sophisticated predictive models for large scale 5G network deployments.