The move from drug candidate through preclinical, then clinical development is fraught with hazards relating to efficacy and adverse drug reactions. Since this aspect of drug development has a high failure rate, much effort is being expended in making the relevant processes more efficient and much less reliant on animal models. The nature of some of these efforts is the subject of this chapter which reviews stem cells, organoids, microphysiology and safety assessment, all supported by machine learning. New approaches to clinical trial design and reporting are covered, with the chapter concluding with short sections on drug delivery and diagnostics.