The Semantic Web vision is among others to allow automatic identification and selection of Web documents and services to meet the requirements of users. In this work we provide a novel solution that supports organizational knowledge flow utilizing non-functional qualitative criteria for web service consumption. A knowledge management web service selection mediator is presented based on the Web Services resource framework (WSRF). It enhances organizational best practices and promotes reusability of successful services during the process of online Web Service selection. Apart from meeting functionality requirements, the mediator utilizes previous domain knowledge to base its decision upon Quality of Web Service (QoWS) ontology knowledge. The proposed solution is open and able to host a number of different selection policies and business logic implementations. We present and experimentally evaluate and compare four such selection policies. The design of the proposed mechanism is analyzed and implementation details are discussed. Evaluation results have shown that the solution is able to satisfy cases and scenarios that have been derived while studying and working on web services selection business processes for different enterprises.