Adoption rates of traditional Operations Research (OR) based decision support systems (DSS) suffer from perceived complexity of the underlying model and its detrimental effect on user-friendliness. The mental effort required to understand abstract models can hinder adoption. This barrier may seem even greater to people with low analytic capabilities. Unfortunately it is often this user group that could benefit the most from using OR based DSS. Agent based approaches on the other hand typically model negotiations between real-world counterparts. Extending cognitive fit theory we argue that presenting DSS in an agent based fashion allows for a closer match between the model presented on screen and the mental model of the user. We tested the impact of DSS presentation on perceived usefulness in a lab experiment (n=118). Our data suggests that an agent presentation outperforms an OR based DSS for perceived usefulness for low analytic users.