Relating, connecting and navigating between concepts represent a major challenge for machine intelligence. On the other hand, collaborative repositories provide a large base of knowledge already filtered, structured, linked and meaningful from a human semantic point of view. Although these repositories are machine accessible, they have no formal explicit semantic tagging to help for automatic navigation in them. In this paper we present a randomized approach, based on Heuristic Semantic Walk (HSW) for searching a collaborative network in order to extract meaningful semantic chains between concepts. The method is based on the use of heuristics defined on semantic proximity measures, which can be easily computed from general search engines statistics. Information from multiple random chains can be used to compute semantic distances between the concepts, as well as to determine the underlying semantic context. The proposed method solves major issues posed by collaborative networks, such as large dimensions, high connectivity degree and dynamical evolution of online networks, which make classical search methods inefficient and unfeasible. In this study the HSW model has been experimented on Wikipedia. Tests held with the well known Word Sym353 benchmark for human evaluation show that the proposed model is comparable to best state-of-the-art results, while being the only web-based approach. Other potential applications range from query expansion, argumentation mining, and simulation of user navigation.