The unconscious application of sophisticated tools, and in particular the popular reverence to data as the source of knowledge, seems to be the rule in many scientific activities, in which the application of tools replaces thinking and data analysis replaces understanding. In this respect, system dynamics has much to offer. One of its trademarks is known as 'operational thinking'. This paper demarcates operational thinking as a distinct epistemic posture that, unlike popular scientific practice that seeks to explain the world by means of data analysis, intends to understand the world in terms of its operations. I explore the significance of such a posture for the domain of human systems and its epistemic value in particular with respect to the prevalent observational approach to science, the Humean problem of induction and determinism. Operational thinking means to recognize that human systems do not obey laws to be discovered by data-analysis, instead, it acknowledges agency, that is, the fact that a social system is the result of the consequences of actions taken by free decision-makers. Operational thinking fits these ontic commitments by providing intelligible explanations that acknowledge responsibility and freedom as characteristic attributes of human systems, whose redesign requires then the capacity to recognize and transform operational arrangements, that is, the intervention in the very decision processes that the systems' constituents carry out.