BT is responsible for continuous decision-making within a state.

It uses BT Root as its single entry point and is typically entered from Flow or FSM. Once started, BTRoot maintains its own tick loop during execution and returns control to the outer layer through Success or Failure. BT is well-suited for priority-based selection, interruption, condition monitoring, and behavior composition. For example, it can express logic such as: attack if an enemy is present; otherwise patrol; otherwise stay idle. In other words, BT is not meant for high-level state switching, but for building the AI decision tree inside a state.