Efficient design of manufacturing networks is paramount for a sustainable growth. The establishment of mass customization and the transition to personalization complicates design activities and leads to vast amounts of unexploited data. This research work aims to exploit existing knowledge for enhancing decision-making during the initial manufacturing networks design, which carry out custom orders of industrial equipment. A method developed into software is proposed, comprising a Genetic Algorithm with knowledge-enriched operators and an intelligent initialization algorithm that exploits existing planning knowledge. The validation of the method is performed using data from a high-precision mold-making manufacturer and its network of first-tier suppliers.