Paper-dependent companies spend most of their time organising and searching for documents, such as invoices, contracts, and budgets. To carry out those tasks, fiduciaries, insurance brokers, and other companies are pioneering Document Management Systems (DMSs). However, there is currently no DMS that allows the classification, understanding, and reasoning of a bundle of documents delivered by customers, and that helps companies create customer profiles to make better and faster decisions. Our proposal aims at easing the tasks of companies dealing with administrative documents of different kinds. We propose a semantic rule-based approach that permits to recognise and classify customers’ documents, as well as to reason over those documents and create customer profiles based on the extracted information. We provide and discuss a case study grounded on the Swiss tax declaration. We developed a full Swiss tax ontology composed of 241 classes, as well as 120 semantic reasoning rules fully validated on the minimum set of administrative documents necessary to fill the Swiss tax declarations. Our approach automates activities related to the management of administrative documents, the profile of their clients, and the administrative documents they must deliver.