Purpose -The purpose of this paper is to describe new search methods based on semantic information structures to offer a personalised and interactive knowledge retrieval to the user. Design/methodology/approach -A careful look on existing "knowledge management projects" leads to a proposed approach of a "knowledge broker network" where transfer of knowledge is based on communication between humans. To enhance the quality of search results a new search algorithm ("Search Assistance Algorithm (SAA)") is described in a simplified manner.Findings -The paper provides information on the knowledge broker network and shows the advantages of this concept especially of the automated part.Research limitations/implications -The paper shows the proposed approach has not formally being tested yet. Some pilot studies have been made and a study especially for the SAA is planned for the near future. Practical implications -This approach has the potential to change the way of transfer of knowledge using a knowledge broker network based on knowledge management systems, which could be improved by our SAA in practice. Originality/value -Using knowledge management systems in a "standard way" by feeding in information and using retrieval methods is not optimal for sharing knowledge between humans. To provide and update information is very cost-intensive. Using "meta-information" about "who is expert in which area" and bring information-seeker and experts together could be the optimal method. This knowledge broker network helps to bring this idea into reality.