An important challenge in CogInfoCom is how to design multi-sensory signals which are capable of communicating high-level information to users. Although there have been investigations on how to design groups of signals for this purpose (referred to as CogInfoCom channels), and how to map high-level concepts onto patterns of sensory signals across different modalities, one general question has still not been addressed: how should a system decide when to provide information to users in interactive CogInfoCom systems? This paper represents an attempt to bridge this gap, and to provide a taxonomy of CogInfoCom trigger types in order to help answer this question. © 2012 IEEE.