To accelerate the drug-discovery and -development stages, one of solutions is to shorten the term of each task contained in the stages. AI has potential to solve this issue and has been applied to several tasks. Furthermore, some AI systems are in development. I will introduce three examples in the drug-discovery stage and one in the development stage, 1: Prediction of human cardiotoxicity, 2: Phenotype screening derived from cell morphological image features, 3: ADMET prediction for the designed compounds, 4: Efficiency improvement of clinical trials collaborating with Hitachi. Hitachi will talk about the fourth one in detail. The aim of these examples is to reduce time-consuming works, to compensate experimental data with highly accurate predictions, or to possess better observing-eyes instead of human eyes.