People are able to determine whether a given document is interesting just by glancing through it. However, when asked to make explicit the rules upon which such a decision is based, they are unable to do so. This, I argue, is because tacit knowledge is involved in this process. Tacit knowledge constitutes the major part of the body of knowledge and it is therefore important for organisations to sustain and exploit this asset. While studying how an intranet recommender system prototype was used, I discovered how tacit knowledge in the form of professional interests could be shared among the organisational members in an unobtrusive way. Based on these empirical findings, and informed by Polanyi's theory of tacit knowledge, I claim that agent-based retrieval systems can be used to capture and visualise our professional interests, thus making otherwise elusive tacit knowledge tangible so that it can benefit others.