The core-mantle interface is the central cog in the Earth's titanic heat engine. As the boundary between the two major convecting parts of the Earth system (the solid silicate mantle and the liquid iron outer core) the properties of this region have a profound influence on the thermochemical and dynamic evolution of the entire planet, including tectonic phenomena at the surface. Evidence from seismology shows that D" (the lowermost few hundred kilometres of the mantle) is strongly heterogeneous in temperature, chemistry, structure and dynamics; this may dominate the long term evolution of the Earth's magnetic field and the morphology of mantle convection and chemical stratification, for example. Mapping and characterising this heterogeneity requires a detailed knowledge of the properties of the constituents and dynamics of D"; this is achievable by resolving its seismic anisotropy. The observation of anisotropy in the shallow lithosphere was an important piece of evidence for the theory of plate tectonics; now such a breakthrough is possible for the analogous deep boundary. We are at a critical juncture where developments in modelling strain in the mantle, petrofabrics and seismic wave propagation can be combined to produce a new generation of integrated models of D", embodying more complete information than any currently available. I propose a groundbreaking project to build such multidisciplinary models and to produce the first complete image of lowermost mantle anisotropy using the best available global, high resolution seismic dataset. The comparison of the models with these data is the key to making a fundamental improvement in our understanding of the thermodynamics and composition of the core-mantle interface, and illuminating its role in the wider Earth system.