Haulage is one of the most significant cost drivers and productivity levers in open‑pit mining. Even modest improvements in truck–shovel interactions can generate substantial operational gains. Yet most planning processes still rely on static assumptions that fail to reflect the variability and constraints of real‑world haulage systems. As a result, mine plans are often difficult for operations teams to execute, and the disconnect between planning and day‑to‑day fleet control can lead to inefficiencies, bottlenecks, and divergence from intended outcomes. Traditional schedules typically assume a fixed dig rate for each shovel and derive truck requirements only as a post‑processing ‘check’. These simplified methods overlook the effect of queuing, delays, routing choices, and the dynamic interactions that shape actual haulage performance. Compounding this challenge, the fleet assignment or autonomous haulage systems used in operations are rarely ‘plan‑aware’. Plans are often communicated through complex documents that do not align with the input formats controllers require, whose decisions and experience significantly influence productivity. This paper introduces a practical, integrated approach to bridging this long‑standing gap between planning and operations. By combining Discrete Event Simulation (DES) with Artificial Intelligence (AI), planners and can evaluate haulage strategies in a realistic virtual environment and identify optimal configurations before passing them to operations. A DES model (Figure 1) replicates the natural variability of truck–shovel operations, incorporating shift constraints, equipment availability, road networks, and expected delays. An AI optimisation engine then explores hundreds of potential haulage scenarios, systematically adjusting parameters to find solutions that best satisfy production, efficiency, and cost objectives.