In recent years, the ease of adaptation of BIM technology was difficult although it brought significant importance and contribution in AEC industry. Different areas use BIM technology however workload necessary to integrate such technology are time-consuming and not easy to learn as BIM practices needs user to bear a lot of field experience. Some of the major challenges to this technology adaptation relates to the classification and analysis aspects of 2D CAD drawings to be converted into a 3D BIM model. Having said that, the article proposed a direct and automated solution with the utilization of machine learning concept to ease the labor-intensive conversion process. With machine learning model’s intelligence to correctly identify different entities in a 2D drawing, the classified information can be built to the visualization process of 3D BIM model. As additional uniqueness to the project, a predictive maintenance analysis with a different machine learning model was explored. The evaluation of the project was based on the accuracy of the neural network in indicating the building elements for visualization on the Autodesk Revit.