Open Source Software (OSS) components are characterised by heterogeneous licenses that give the possibility to use, modify and of-Ten redistribute the source code. Their adoption meets several adopter's needs, such as cost reduction, standards alignment, and so on. However, often OSS projects retain several different (or missing) licenses for the various components and files, which raise risks of violations and potential legal issues, if not correctly managed. This makes necessary to understand the characteristics and implications of licensing and their relation to the adopter's goals. In this paper we report the use of risk assessment techniques to make inference about license risk exposure associated with each business goal. We rely on existing knowledge, gathered from domain experts, and map it onto formal models that can be automatically analysed to provide some evidence about relevant license in- formation and related risk. Goals are used to drive the software license selection. We illustrate the approach for the case of a research and innovation action project funded under the H2020 framework.