In auto-chemotaxis agents (cells, micro-organisms, ...) move towards high concentrations of a pheromone (signalling chemical) that they produce themselves. This leads to a delayed and long-range coupling between agents and intriguing collective phenomena. We clarify the kind of cooperative behaviours that enable large populations to establish and follow trails for long distances by studying hydrodynamic limits of agent-based models using hybrid simulations and stochastic process convergence theory.