In this research, we present a three-dimensional infrastructure data platform developed by the authors. The study highlights that three-dimensional models, as opposed to two-dimensional ones, hold a greater quantity of information, enabling a diverse range of analyses. Particularly, this paper delves into applications stemming from the data of the three-dimensional infrastructure data platform, including image processing AI for damage detection, text explanation output, and the automated generation of Finite Element Method (FEM) models. The image processing AI initially identifies the locations of cracks and corrosion damage within captured images, which are then used to construct a plotted three-dimensional model. Additionally, to facilitate comprehension by junior technicians and infrastructure managers who may not possess specialized knowledge in damage assessment, an AI has been developed using image captioning technology. This AI outputs explanatory texts about the detected damage. Furthermore, these capabilities have been integrated into a web system, enabling access to the AI via mobile devices such as tablets at infrastructure inspection sites. This system includes a self-learning feature that improves its intelligence the more it is used, and its accuracy is continually being enhanced. Regarding the automated generation of FEM models, the AI constructs these models with necessary precision using dimensions derived from drawings or point cloud data. This allows for a more comprehensive diagnosis of infrastructure, including mechanical behavior, which has often been overlooked in traditional inspections and diagnoses. The paper also discusses how these AI models are centrally managed on a Cesium-based visualization platform, and the results of this integration are presented. The study showcases how this approach significantly advances the field of infrastructure inspection and diagnostics, leveraging AI and machine learning to enhance accuracy and efficiency.