Many business scenarios require humans to interact with workflows. To support humans as unobtrusively as possible in the execution of their activities, it is important to keep the interaction time experienced by humans as low as possible. The time required for such interactions is influenced by two factors: First, by the runtime of the services that are used by a workflow during an interaction. Second, by the time required to transfer data between workflow servers and services that may be distributed in a global network. We propose an algorithm that computes a suitable distribution of a workflow in such a network. The goal of our algorithm is to minimize the time required for interactions between a human and a workflow. Current approaches in the domain of workflow optimization pay little attention towards optimizing a workflow to increase the usability for humans. We show the feasibility of our approach by comparing our algorithm with two non-distributed approaches and a distributed approach which is based on a greedy algorithm and show that our algorithm outperforms these approaches.