Linkage synthesis involves developing linkages that fulfil the requirements specified for a process or other technical devices. Traditionally, linkage synthesis is carried out in two steps: firstly, structural synthesis, in which the linkage topology, i.e. the number of links and joints as well as joint types and their arrangement in relation to each other, is defined, followed by the dimensional synthesis, in which kinematic parameters such as link lengths and joint coordinates are determined. Both synthesis steps represent a challenge in their own right: There are no direct mathematical methods for topological synthesis that generate the “correct” topology for given requirements, so this is done based on empirical values and heuristic methods. The authors assume that the complete synthesis process can be carried out by a single AI system, provided that a generally valid data set of as many different linkages as possible, including their detailed analysis, is available. Furthermore, it will be shown that, as a replacement for the classic dichotomy of the synthesis process, a holistic approach offers a wide range of advantages.