Safety indicators are important for safety management in any industrial sector. There is growing interest in leading safety indicators to complement lagging safety indicators. This is indicative of a shift towards proactive safety management. But how can we identify and develop leading safety indicators that are meaningfully related to practice? In this paper we describe an approach to the development of leading indicators based on the Functional Resonance Analysis Method (FRAM). The FRAM is an increasingly popular method to analyse complex systems from a Resilience Engineering perspective. Starting from a functional representation of the work domain, the FRAM focuses on modelling potentially non-linear couplings between functions in order to better understand the complexity of the system. We illustrate the proposed approach through a case study of hazardous materials transfer between road tankers and bulk storage facilities.