AGN belong to the Universe's most powerful non-explosive sources, and their emission covers the whole electro-magnetic spectrum. Different methods are therefore employed to identify AGN, but, most importantly, the various wavelength regimes provide different windows on AGN physics. The physical processes that may impact whether an AGN is detected in a given astronomical observation at a given wavelength and which class it is then assigned to may include orientation, accretion rate, black hole mass, environment, host galaxy properties, and cosmic time (redshift). Understanding the differences and similarities between these different AGN classes is a pre-requisite to understanding AGN physics and the role AGN play in galaxy evolution. We have compiled a sample of 632 AGN among the galaxies observed within the large MaNGA IFU survey. These AGN were identified through independent selection techniques, such as radio, BPT, broad-lines, mid-IR color, and X-ray selections. Using the MaNGA IFU data, we have mapped and characterized ionized gas flows in these sources to understand the dependence of galaxy-wide outflows on host galaxy properties, AGN class, and AGN luminosity. I will present how wind characteristics and outflow sizes differ between AGN samples, which parameters drive these differences, and how strongly AGN feedback signatures suffer from selection bias. I will also show that radio-selected AGN show stronger outflows and larger outflow sizes when carefully matched to control samples which suggests that radio-selected AGN may trace an evolved phase of AGN activity. Our results are important in the context of AGN duty cycle and highlight IFU data's potential to deepen our knowledge of AGN and galaxy evolution.