Cultures of hybridoma cells in bioreactors are commonly used to produce monoclonal antibodies. As an alternative to nonlinear model predictive control, which has recently been applied successfully to optimize the process productivity, a simpler control approach is developed in the present study. This strategy is based on a classical PI controller and software sensors for the reaction rates, which exploit the particular structure of the dynamic model. The dynamic behavior of the process can indeed be subdivided into four operating zones, depending on the overflow metabolism of the hybridoma cells. In addition, robustness toward model uncertainties and measurement noise is investigated.