In this paper, we investigate 2727 galaxies observed by MaNGA as of June 2016 to develop spatially resolved techniques for identifying signatures of active galactic nuclei (AGN). We identify 303 AGN candidates. The additional spatial dimension imposes challenges in identifying AGN due to contamination from diffuse ionized gas, extra-planar gas and photoionization by hot stars. We show that the combination of spatially-resolved line diagnostic diagrams and additional cuts on H$\alpha$ surface brighness and H$\alpha$ equivalent width can distinguish between AGN-like signatures and high-metallicity galaxies with LINER-like spectra. Low mass galaxies with high specific star formation rates are particularly difficult to diagnose and routinely show diagnostic line ratios outside of the standard star-formation locus. We develop a new diagnostic -- the distance from the standard diagnostic line in the line-ratios space -- to evaluate the significance of the deviation from the star-formation locus. We find 173 galaxies that would not have been selected as AGN candidates based on single-fibre spectral measurements but exhibit photoionization signatures suggestive of AGN activity in the MaNGA resolved observations, underscoring the power of large integral field unit (IFU) surveys. A complete census of these new AGN candidates is necessary to understand their nature and probe the complex co-evolution of supermassive black holes and their hosts.