The Société de transport de Montréal (STM) is committed to fully electrifying its vehicle fleet as part of its long-term sustainability strategy. This study aims to determine the required number of electric service vehicles and their optimal allocation across STM depots to ensure efficient incident response throughout the bus network. These vehicles are operated by Operations Supervisors, who travel across the network to resolve equipment failures, support drivers, manage passenger-related incidents, and ultimately reduce service disruptions. To address this challenge, we develop a lexicographic bi-objective mixed-integer linear programming model that prioritizes minimizing the fleet size before reducing total response time. The model is tested on small instances based on incident data and solved using CPLEX. A preliminary decomposition approach shows potential for improving scalability. Preliminary results demonstrate encouraging performance.