During the past decade, the pharmaceutical industry has invested considerably in technologies that have the potential to increase throughput in discovery projects. For large compound libraries, efficacy, availability and safety should be determined as early and as reliably as possible. The latest step in this effort is the implementation of in silico methods that combine and interpret (sometimes replace) experimental in vitro data. For ADME properties (absorption, distribution, metabolism and excretion) rational predictive models have been developed that rely on basic physicochemical input data and on mechanistic descriptions of the underlying biophysical and biochemical processes. Some of these models have become commercially available (e.g., GastroPlus: Simulations Plus; PK-Map, PK-Sim: Bayer Technology Services). The contribution of such models to an optimised research and development process will be discussed.