The integration of artificial intelligence (AI) into autonomous excavation systems offers promising advancements in operational efficiency, adaptability, and automation. Our research will explore AI methodologies for enhancing high-level decision-making - leverage AI to optimize strategic decisions, such as determining bucket paths and adjusting to varying environmental conditions in real time, creating a more flexible and efficient system. Building on recent developments in AI-driven control frameworks, including reinforcement learning and machine learning approaches, we aim to implement adaptive AI algorithms that improve the overall excavation process. AI will assist with real-time decisions, by selecting and executing the best possible action on the excavation task. This work primarily focuses on AI integration into the highest level of system coordination - decision-making - while using existing control methodologies at the lower levels, introducing AI as a helpful tool and not a full control solution. At this stage, our approach is conceptual, aiming to establish a scalable framework that can be adapted across various applications, from construction to mining. By utilizing the Isaac Sim platform for simulation, we will model and test the proposed control system under various conditions, ensuring its viability in diverse scenarios. Final results will be verified on an excavator prototype, allowing for practical validation of the AI-based decision-making system and its real-world applications. We hope to drive further exploration into how autonomous systems can not only automate but also optimize complex excavation tasks, ultimately setting the stage for more intelligent, responsive machines.