With “datafication” practices becoming more common in digital ecologies, humans have become increasingly reliant on emerging technologies and other actors that can store, comprehend, and analyze information. This thesis offers a proposed model of mediative agency to address the importance of interrogating how non-human actors interpret and make meaning from data. Mediative agents contribute to the disbursement of rhetoric, as well as our understanding of information, by granting visibility and assigning value to data. These processes effectively play a role in shaping reality through agents’ parameterization of data broadly, allowing non-human actors to take on a complex agency that can alter rhetorical trajectories. In interrogating the structures of power that contribute to the dissemination of rhetoric within digital ecologies, mediative agency acts as a speculative modeling approach that allows rhetors to theorize various functions of agency and anticipate how non-human agency might further develop as technological environments change in the immediate future.