Considering the variety of architectural Cultural Heritage typologies, systemic architectures require specific attention in the recovery process. The dimensions of "extension" and "recurrence" at geographic and technological levels affect the complexity of their knowledge process; they require systematic ways for their categorisation and comprehension to guarantee correct diagnosis and suitable rehabilitation. Recent applications involving Internet of Things (IoT) for the built Cultural Heritage have demonstrated the potentialities of three-dimensional (3D) geographic information system (GIS) models and structured databases in supporting complex degrees of knowledge for technicians, as well as management for administrators. Starting from such experiences, the work presents the setting up of a web-based platform to support the knowledge and management of systemic architectures, considering the geographical distribution of fabrics, natural and anthropic boundary conditions, and technical and administrative details. The platform takes advantage of digital models, machine and deep learning procedures and relational databases, in a GIS-based environment, for the recognition and categorisation of prevalent physical and qualitative features of systemic architectures, the recognition and qualification of dominant and recurrent decays and the management of recovery activities in a semi-automatic way. Specifically, the main digital objects used for testing the applied techniques and setting up the platform are based on Red-Green-Blue (RGB) and mapped point clouds of the historical Telegraphic Towers located along the Madrid-Valencia path, resulting from the on-site investigations. Their choice is motivated by the high level of knowledge about the cases reached in the last years by the authors, allowing them to test rules within the decision support systems and innovative techniques for their decay mapping. As the experience has demonstrated, the systematisation of technical details and operative pipeline of methods and tools allow the normalisation and standardisation of the intervention selection process; this offers policymakers an innovative tool based on traditional procedures for conservation plans, coherent with a priority-based practice.