The objective of the proposed research is to study how to support human-agent collaboration for decision-making by studying Shared Mental Models in hierarchical teams. Humans collaborate with various forms of automation because of its ability to augment performance and offload workload. Highly effective human teams have a mutual understanding of the external world, tasks, responsibilities, capabilities, and limits shared by team members, known as Shared Mental Models (SMM). This theory has been extended to human-AI teams (HAT)s for achieving effective performance. However, the development, maintenance, and utilization of SMMs have been studied within homogeneous dyads where both the human and AI share the goals, roles, and responsibilities, and act upon common information. Therefore, the ensuing team dynamics are simple. As the scale of AI deployment grows, humans will be uniquely poised to engage with multiple agents. They will transition from merely collaborating to supervising and managing the activities of AI teams. A hierarchical reorganization of activity between humans and AI will lead to a smoother transition to multi-agent teaming and execution of tasks. However, in realistic hierarchical teams, the distribution of tasks and information between members of the team, their interdependencies, communication patterns, etc. will bring upon different constraints in the way the human interprets and works with their AI partners. This leads us to the crux of our proposed thesis- to study how hierarchical teaming between humans and AI impacts SMM formation in a geospatial decision-making context. We seek to support the human in the HAT by focusing on the lower stability aspects of their mental models i.e., their understanding of the teammates. We aim to identify the factors affecting the development and maintenance of SMMs between humans and their AI teammates and how the overall team operations get affected with the human operating at the helm of the team hierarchy.