This article studies the cash distribution problem for bank correspondents that are geographically dispersed, and the associated risk indicators, service times, and logistics costs. The proposed model is a variant of the Inventory Routing Problem (IRP) that considers risk-management strategies. The model considers a single cash center with unlimited capacity, from which cash deliveries are carried out by a single vehicle with limited capacity for cash distribution. The bank correspondents to which the cash deliveries are made handles a limited storage capacity. This paper considers the routing decisions of vehicles with hard time window, and risk constraints focused on inducing unpredictability of routes in a cash supply chain. The proposed model is denoted as a risk constrained inventory routing problem with time Windows (RcIRPTW). A mathematical formulation based on mixed-integer programming is proposed, studying the impact of incidents risks in the transport of cash, generating computational experiments on test instances where four types of variations of the IRP are evaluated with several random demand scenarios. The results show the importance of simultaneously optimizing inventory decisions considering routing and inventory costs along with the consideration of risk mitigation strategies. These risk constraints force changing the routes of the vehicle, reducing incidents rates, and reducing route predictability for secure transportation and logistics.