We describe an AI behavior architecture that supports responsive collaborative interruptible and resumable behaviors using behavior queues. This architecture wraps sets of behaviors into roles, which provide a simple efficient mechanism for encapsulating behaviors into components that can change dynamically, based on environmental criteria. To illustrate the viability of this architecture in a commercial setting, we implemented this architecture in BioWare Corp.'s Neverwinter Nights game. To demonstrate the usability of this architecture by game designers, we implemented a simple interface in ScriptEase, so that the architecture can be utilized with no coding skills.