Earth modeling plays a decisive role in seismic imaging. Presently, methods such as tomography and full-waveform inversion (FWI) are widely used to generate 3D high-resolution velocity models across sedimentary basins. However, given the nonunique nature of the solution, the relative fidelity of these velocity models remains low. Moreover, earth-model building must incorporate a dual strategy. In particular, it must derive a velocity model for migration and yield a series of rock properties such as density, temperature, effective stress, and pore pressure. These large-scale, basin-sized rock properties are used for drilling purposes and as an initial model for small-scale reservoir property inversion. Hence, one must inquire: How can we improve an earth model built at basin-scale using seismic data? Traditional workflows will simply convert seismic-derived velocity into rock properties, thereby propagating the uncertainty without addressing the issue. To alleviate this problem and fully exploit the potential of seismic data, an alternate workflow will be discussed. The new workflow involves using rock physics to link rock properties with velocity, and the physical range of the rock properties is used to constrain velocity when derived from surface seismic data; this way we reinforce the reliability of the final earth model. The application of this new workflow is demonstrated using Gulf of Mexico examples.