Maintaining clean data in organizations is crucial and strategic. It has been indeed proven that poor quality data can lead to efficiency and/or financial damages. For this purpose, organizations are now relying on the so-called Master Data Management systems. These systems aim at providing organizations with tools for building robust sets of information acting as the truth. They are for instance used for maintaining up-to-date information about the entities, customers and products of a company. Many tools have recently been proposed for managing master data, but they do not allow to consider the three key points we address in our work, namely genericity, graph-orientation and collaborative solutions. In this paper, we present the KEBENARAN system that integrates all these three concepts. We show the feasibility of such an approach through the implementation of a prototype.