Mining haul trucks are responsible for up to 75 per cent of mine site emissions, posing a critical challenge for decarbonising open pit operations. While zero-emission technologies offer long-term solutions, they remain economically and logistically constrained. This study introduces an Al- powered optimisation framework integrating high-precision fuel monitoring, real-time diagnostics, and machine learning analytics to improve fuel efficiency. Deployment across four mine sites reduced fuel consumption by 6.2 per cent to 10.3 per cent, surpassing Australia's 4.9 per cent CO2 reduction mandate. Key interventions included operator feedback systems, predictive maintenance, and haul route optimisation. The system demonstrated a return on investment within six months and integrated seamlessly with existing fleet management software. These findings underscore the framework's scalability, cost-effectiveness, and alignment with sustainability and operational goals, positioning it as a viable bridge to zero-emission haulage.