A decision support model aimed at the assessment of reputational risks associated with bank operation is presented in this paper. The innovative aspect of the model is that it combines different types and sources of information: structured and unstructured, quantitative and qualitative, internal and external. Unstructured, qualitative and external aspects are represented by sentiment of news and blogs about bank counterpart organisations. The model is multi-attribute and hierarchical, and is composed of three modules: basic data processing, qualitative evaluation, aggregation; these are described in detail. This paper also presents a prototype implementation of the model, illustrates its application on real-life data, and reports on experts' opinion about the model and its use.