The application of artificial intelligences on digital games became more and more successful in recent years. A drawback is, that they need lots of computing power to achieve good results, the more complex the game, the more computing power is needed. In this thesis a strategy learning-system is implemented, which is based on crowd-learned heuristics. The heuristics are given in a wiki. The research is done according to the Design Science Research Methodology. The implemented system is allied to the game Dominion. To do this, an ontology for Dominion is designed. A mapping language is defined and implemented in the system, which allows the mapping of information in the wiki to an ontology. Furthermore, metrics to rate the found strategies are defined. Using the system, users can enter a mapping for the information transfer and apply it. They can also select cards from Dominion, for which the system determines and rates strategies. Finally, the system is evaluated by Dominion-players by rating the strategies, which are found by the system, and the defined metrics.