The continued growth of air traffic demand is placing increasing pressure on current air traffic control (ATC) systems, prompting the need for alternative ATC strategies. A promising approach is a shared ATC environment between a human controller and an automated controller, where basic, low-complexity traffic is delegated to automation while complex traffic remains under human control. This concept requires a reliable method for predicting the operational complexity of individual flights. This research presents the design and evaluation of a complexity-based flight allocation algorithm for an en-route shared human–automation ATC environment. The allocator classifies incoming aircraft based primarily on the predicted number of interactions along their trajectories, using a flight-filtering mechanism derived from existing models. Additional allocation metrics include the expected number of interactions between human and automation-directed flights and a minimum number of flights controlled by each controller. The allocator was evaluated using offline simulations with real traffic data, followed by a human-in-the-loop experiment with two professional air traffic controllers. Results show that the allocator can consistently assign more complex flights to the human controller while maintaining a balanced workload distribution. The human-in-the-loop experiment saw substantial manual re-allocation and revealed low trust in both the allocator and the automation, indicating the need for further refinement and closer integration with automation capabilities.