The use of unstructured data in (re)insurance has become a central topic with the arrival of new statistical practices related to data science. Today, we have many examples that prove there is a growing interest in gathering and exploiting these data: optimizing internal processes, improving risks understanding, getting an overview of its market or regulations, etc. Reinsurance treaties are essential documents within the reinsurance value chain. These documents exist to provide a contractual framework between reinsurers and ceding companies. They are also full of valuable information that are sometimes collected manually, but in most cases are simply left unexploited. The aim of this presentation is to propose a set of methods that allows to analyze and to compare reinsurance treaties automatically. Apart from methodological and IT aspects, the study will also present several innovation enablers that allow to enhance the generated data in actuarial frameworks such as underwriting, pricing or risk management.