The timing is deliberate. ICAO’s Air Navigation World conference opens in Montreal today with air-traffic-flow management at its centre, as states prepare for new global ATFM provisions due to apply in November 2030.
SMART (Strategic Management of Airspace, Routes and Trajectories) is not an autonomous controller. It is a decision-support system. Software that combines about 200 data streams, including weather, flight paths, airline schedules, airport capacity, traffic-flow conditions and controller staffing, then identifies where demand may exceed the system’s ability to handle it.
The FAA has begun limited use of the tool in the Washington area, covering Ronald Reagan Washington National, Washington Dulles and Baltimore/Washington International. Air Space Intelligence is building the platform under a 12-year contract worth up to $875 million, with the FAA intending a staged national expansion.
In practice, SMART’s purpose is to answer questions that flow managers currently answer with several disconnected systems, forecasts and phone calls. If thunderstorms close a preferred arrival route in four hours which flights will converge on the remaining routes. Whether Dulles, Reagan or BWI will become the bottleneck and whether holding aircraft on the ground at their departure airports would be less disruptive than having them queue in the air.
A small disruption becomes expensive when it spreads. A late inbound aircraft misses its turnaround. The delayed crew approaches duty limits. The next flight loses its departure slot; arriving aircraft must hold because gates remain occupied. Then the delay reaches airports far from the original weather cell.
SMART seeks to model that chain before it fully develops. If it sees that forecast weather, demand and staffing will overload an arrival bank, it can present controllers and traffic-management specialists with options, hold selected flights at the gate, issue a ground-delay programme, stagger departures, reroute flights, adjust arrival rates, or prioritise flights according to agreed operational rules.
The important qualifier is that SMART recommends and people decide. Air-traffic managers still have to assess safety, fairness, fuel, passenger connections, crew duty limits, airline priorities and the possibility that a weather forecast will be wrong. A predictive system can improve the quality and timing of those decisions, but it cannot remove the trade-off inherent in constrained airspace.
The Washington region is a demanding test case because Reagan, Dulles and BWI share crowded regional airspace while serving different traffic patterns. Reagan is constrained by its location and runway system, Dulles carries substantial domestic and international traffic, and BWI is a major low-cost carrier base. A disruption at one can affect routes, controllers and airport capacity across the wider network.
The first deployment is still a trial. The FAA says it is operating SMART in a limited mode and intends to expand gradually, rather than handing network control to an AI model.
ICAO’s Montreal meeting is not about one US product. It is about making air-traffic-flow management a more consistent global capability. The October 5-9 programme is intended to give states and air-navigation service providers practical guidance as they prepare for new ICAO ATFM provisions planned for 2030.
AI may make that process faster and more granular. It can merge data that previously sat in separate weather, flight-plan, staffing and airport systems. But ICAO’s challenge is governance as much as technology: common data definitions, cyber security, explainable recommendations, human accountability and equitable treatment for airlines from different states.
India already has a form of flow management at its busiest airports. Airports Authority of India says its Future India Air Navigation programme includes air-traffic-flow management over busy routes, expanded automation at non-metro control towers and GAGAN satellite-based augmentation.
Washington is testing whether AI can make that shift operationally useful. Montreal is asking whether the world’s air-navigation systems can build the common rules, data and human oversight needed to use that capability safely.