Abstract A future challenge for the European air traffic system is an estimated rise in traffic load in combination with the occurrence of so-called disruptive events (e.g., thunderstorms, security alarms or union strikes). EUROCONTROL showed in their simulations that external events lead to instabilities in the whole European air traffic system with a weakened ability to recover [cf. (EUROCONTROL, Challenges of growth 2013—Task 6: the effect of air traffic network congestion 2035. EUROCONTROL, Brussels, 2013)]. Due to these events, EUROCONTROL estimates a strong increase of delay within the air traffic system, especially at the airports. Among various factors to keep the impact of these events minimized, a sufficient cooperation between the affected air traffic participants (pilots, controllers, airport staff, etc.) is necessary. Sharing information and aligning on a procedure to utilize the existing resources offers a way to efficiently keep up operations. A tool to train this cooperative behavior and to validate the benefits of new systems and procedures regarding the enforcement of cooperative behavior is a human-in-the-loop simulation (HITL). HITL simulations enable cooperative situations by facing the participants with disruptive events. These situations are commonly supported by so-called pseudo-actors, humans which take on one or multiple roles within the cooperation. Thereby the cooperation can be developed in a more or less challenging way. For instance, a delayed pilot reaction in dense traffic or an airline dispatcher not willing to cancel flights into a congested airport can be performed by a pseudo-actor. Currently, pseudo-actors with predefined instructions for their behavior are the only way to generate the above-mentioned challenges in HITL simulations. As human beings, the pseudo-actors show varying behavior. Therefore, reproducibility of the HITL simulation is limited, although being one of the key features. In contrast to field trials in a real environment all influencing factors of a HITL simulation are under control. Situations can be repeated multiple times. In terms of training simulation this is necessary to generate the intended learning effect. In terms of validation simulation the possibility to exactly repeat a situation is needed to compare a situation with the support of a new system (experimental condition) to one without the new system (baseline). This paper provides a contribution for assessing the potential of human pseudo-actor alternatives. Virtual agents are already used in other fields such as gaming and other types of simulation (e.g., virtual military training simulation). These virtual agents interact with the other human participants of the simulation. Moreover, as being computer applications, they strictly do follow their programmed behavior. Thus, they offer the benefit of reproducible behavior. The evaluation within this paper uses the airport management simulation of the German Aerospace Center to challenge four participants with multiple disruptive events. The role of the stand and gate allocation manager is either taken by a human pseudo-actor or a virtual agent. The results of the simulation runs with the human pseudo-actor are compared to the ones with the virtual agent. Analyzing the results a first conclusion on the potential of virtual agents in air traffic HITL simulations will be given.