The aircraft are further along than the system that must decide for them. Keeping ever more aircraft safely separated is an architecture problem, not an AI problem. ACT is a unique architecture that embeds a digital twin into live NAS operations, continuously generating, simulating, and scoring options while authority stays with the operator. We see these human-over-the-loop architectures as the key to integrating new-entrant operations at scale.
Digital twin, in the loopA simulation runs continuously alongside live operations, so decisions can be tried safely in the twin before they are made for real.
Architecture first, AI secondAI plugs in only where it is genuinely better, such as predicting traffic and weather or scoring options, and never holds authority over the airspace.
Human over the loop, at scaleOperators keep authority at every decision point. This is the key to bringing drones and new aircraft into the airspace at scale.
AI plugs in at defined points; the human gate is never removed. Notional.
Live system stateWhat is happening now
Prediction engineAI sees ahead
Digital twin / sim coreAI tries options safely
Strategy & scoringAI ranks the choices
Human decision layerThe human decides
Every decision and its conditions are logged, so a strategy can be evaluated after the fact rather than defended from memory. Avansys is explicit with customers about where the build currently stands: a tool built to certify operational readiness cannot be sold on claims its builders would not themselves certify.