Sharing landscapes with wildlife is difficult because these landscapes have the potential to bring multiple costs for both people and wildlife. Finding solutions to promote human-wildlife coexistence is therefore critical, especially in the face of rapid global biodiversity loss. While traditionally the problem was addressed by implementing technological solutions to reduce negative wildlife impacts, more recently the complexity of the problem is recognized but tools and methods to deal with this complexity are scarce. In this perspective we apply causal loop diagrams and a leverage point perspective to synthesize the results of our long-term research in communal conservancies in the Zambezi region of Namibia and to better understand the complexity of this human-wildlife social-ecological system. After identifying 32 variables and 47 relationships between these variables, we constructed two causal loop diagrams consisting of two sub-systems — a “governance” sub-system and a “wildlife” sub-system. From the 32 variables, we identified four that could potentially leverage systemic change to improve human-wildlife coexistence. These were: 1. conflict between people, 2. tolerance by people for wildlife, 3. policy clarity, relevance, and congruence; and 4. congruence and relevance of governance structure with local conditions. Social learning programs involving training in nonviolent communication, global sustainability challenges and new governance models emerged as potentially valuable interventions. Our approach focusing on causal relationships and leverage points was useful to synthesize complex findings from long-term research and engage stakeholders in joint learning processes.