Knowledge graph embeddings and successive machine learning models represent a topic that has been gaining popularity in recent research. These allow the use of graph-structured data for applications that, by definition, rely on numerical feature vectors as inputs. In this context, the transformation of knowledge graphs into sets of numerical feature vectors is performed by embedding algorithms, which map the elements of the graph into a low-dimensional embedding space. However, these methods mostly assume a static knowledge graph, so subsequent updates inevitably require a re-run of the embedding process. In this work the Navi Approach is introduced which aims to maintain advantages of established embedding methods while making them accessible to dynamic domains. Relational Graph Convolutional Networks are adapted for reconstructing node embeddings based solely on local neighborhoods. Moreover, the approach is independent of the original embedding process, as it only considers its resulting embeddings. Preliminary results suggest that the performance of successive machine learning tasks is at least maintained without the need of relearning the embeddings nor the machine learning models. Often, using the reconstructed embeddings instead of the original ones even leads to an increase in performance.