For the big companies, picking up a number of employees from various places (residential areas or common places) within a neighborhood and brought to the factory should be a challenging task. Potential improvements for this constantly repeated service planning may result in cost and total traveled distance reduction. To search and provide the aforementioned potential savings, three different scenarios with three different bus capacities are generated for a textile company in this paper. In the first scenario that is also the current application of the company, an integer programming model for the capacitated vehicle routing problem (CVRP) is applied. In this scenario, shuttle buses with a certain capacity are routed by picking up the employees from their homes. In the second scenario, the employees are clustered (cluster size is equal to the vehicle capacity) using P-median model to minimize the total walking distance of employees to the meeting point, then the shuttle buses are routed by visiting the clustered zones (meeting points). In the last scenario, the employees are clustered using K-means model and then, shuttle buses transport workers to factory from the center points of each cluster. Three scenarios with different bus capacities (20, 30 and 50 people) are applied to a textile company (the depot) that includes 1,361 employees. The models are run using Gurobi 9.1.1 with Python 3.8 and the results are discussed. According to the results, second and third scenarios reduce the total traveled distance by 46.90% and 44.10% (averagely) compared to the first scenario, respectively.