In the context of an agent-based modeling that esti mates the acceptance of telecenters by their potent ial users in Brazil, this paper discusses two alternativ e configurations of neighborhood for the creation o f social networks, one based on Moore’s cellular auto maton, and another based on random relations. This discussion is an essential step in the creation of the overall model, a bottom up approach in which th e users (agents) are characterized in terms of their technological innovativeness and are subject to the influence of their social networks; each agent is r andomly assigned with a set of initial features in order to reflect the diversity of behaviors and preferenc es found in a field survey with the target audience . The final model is expected to be an evaluation tool to compare alternative digital inclusion strategies i n terms of effectiveness in attracting new users to t he telecenters, and assessing the policies in terms of their efficacy in the resource allocations and of h ow effective is a given equipment deployment strate gy or premise location. This can help to reallocate resources so as to maximize the results.