Significance: New chemical entities (NCEs) with desired pharmacological and biological activity spectra fuel drug discovery and provide tools for chemical biologists. Computer-assisted molecular design generates novel chemotypes with predictable polypharmacologies. We present the successful application of fully automated de novo drug design coupled with a pioneering approach for target panel prediction to obtain readily synthesizable bioactive compounds. This innovative concept enabled the identification of relevant macromolecular targets of computationally designed NCEs and led to the discovery of previously unknown off-targets of approved drugs.