Work environments can profoundly influence employee well-being, workload, and performance. As well-being and workload have complex bidirectional interactions, consequently employees can be captured in healthy or unhealthy vicious cycles for their dynamic interplay. Our aim is to model these vicious cycles as competing feedback loops, based on a network modeling perspective from artificial intelligence (AI). In this human-centered AI approach, we specifically focus on contextualization: on how our newly developed context-sensitive competing feedback loops model can capture employee well-being under changing working environments in an organization. We simulate five dynamic contextual factors in the working environment, such as market demand and management support, and evaluate how employee well-being develops over a one-year course. When the working environments became worse, employees were more stuck in an unhealthy burn-out feedback loop, but once there was intervention in some contextual factors, employees switched to the healthy feedback loop again. From these simulations, we conclude that contextualized competing feedback loops can offer a flexible framework to model the dynamics of employee wellbeing and workload depending on working environments. Such contextualized competing feedback loops might serve as a good mental health application tool.