When a flight’s operational conditions change (e.g., an updated weather forecast), various alternative trajectories may be computed. These usually require trade-offs between expected fuel burn and delay. The pilot, or the dispatcher, considers these expected values to decide how to operate the flight. This approach has two main challenges. Firstly, it requires the translation of arrival delay into parameters that are relevant for the airline (on-time performance and cost of delay). Secondly, uncertainties in the system need to be estimated (e.g., holding at arrival). These estimations rely on airline staff expertise. Pilot3 sets out to overcome these issues by developing a new, multi-criteria decision-support tool, which incorporates explicit estimators for performance indicators and ATM operational parameters. These estimators will be developed incrementally, from simple heuristics to advanced machine-learning models, building on previous experience.