NASA is researching various concepts, procedures, standards, and technologies intended for NextGen Airspace. Complex, distributed airspace simulations that utilize experimental testbeds (e.g., Multi Aircraft Control System, or MACS) are vital research tools for these projects. However, managing the various complexities and coordination of agent-supported separation assurance can be challenging. This often creates undesired staffing and training requirements, workload, and susceptibility to human error that can disrupt planned scenario events. To address this issue, we propose to develop Artificial Intelligence for Refining Multi-Aircraft Testbed Environments (AIR-MATE). This proposed innovation will provide a MACS-interoperable software module that coordinates the behaviors of human-automation pairs in simulated NextGen airspace. This effort will leverage recent advancements in distributed constraints optimization and adjustable autonomy to analyze airspace simulations in a decentralized, parallel manner and solve problems locally for enhanced efficiency. This technology will reduce the workload and staffing requirements in current NextGen simulations, while ensuring the desired scenario events and separation assurance is properly executed. The results of the AIR-MATE effort will be a more controlled and high-fidelity testbed environment that will aid researchers, increase the quality of NextGen research, and ultimately benefit the development of NextGen concepts.