Purpose – The purpose of this paper is to focus on the underpinning dynamics that explain collective intelligence. Design/methodology/approach – Collective intelligence can be understood as the capacity of a collective system to evolve toward higher order complexity through networks of individual capacities. The authors observed two collective systems as examples of the dynamic processes of complex networks – the wiki course PeSO at the Universidad de Los Andes, Bogotá, Colombia, and an agent-based model inspired by wiki systems. Findings – The results of the wiki course PeSO and the model are contrasted with a random network baseline model. Both the wiki course and the model show dynamics of accumulation, in which statistical properties of non-equilibrium networks appear. Research limitations/implications – The work is based on network science. The authors analyzed data from two kinds of networks: the wiki course PeSO and an agent-based model. Limitations due to the number of computations and complexity appeared when there was a high order of magnitude of agents. Practical implications – Better understanding can allow for the measurement and design of systems based on collective intelligence. Originality/value – The results show how collective intelligence emerges from cumulative dynamics.