In this paper, we propose a system for automatically optimizing a decision support system to assist IT helpdesk operators for IT management tasks. The proposed system aims to leverage the publicly available experience that can be accessed on the Internet (such as blogs, discussion list etc.) in order to manage problems in IT environments. To achieve this, we propose a data driven system that is based on a combination of clustering and optimization techniques. Our results show that the proposed system can work on real life data sets and is a promising model for supporting decision making processes in IT environments.