In support of exploration programs across the deepwater Gulf of Mexico, prestack depth-migrated (PSDM) seismic data sets have come to be widely used for prospect identification and risk assessment. These projects have often followed the 80% solution, where the standard has been isotropic velocity analysis and imaging. However, with sustained exploration success, there has been an increasing need to improve both the image quality and predictive capability of PSDM data sets in support of project appraisal and development. Achieving the desired improvements in image fidelity places many requirements on our approach to seismic imaging and velocity modeling. Foremost among these is accuracy; the imaging must be able to precisely image faults, subtle stratigraphy, and complex, steeply dipping features. In parallel, our approach to velocity modeling should provide a robust, flexible, and integrated methodology for describing the complexities of anisotropic Earth models, since sharply defining the geologic structure will be critically dependent on accurately honoring the lateral velocity variations and anisotropy that are likely to be present in a high-quality Earth model.