A major component of any training system is the instructor and the expertise he brings to the system. Because instructors are a critical resource often in short supply, automating the instructors' role, or part of that role, especially in real-time simulations is highly desirable. Automating the instructor is difficult because an instructor's knowledge embodies an understanding of a myriad of relationships including complex dynamic relations, as well as reasoning strategies, experimental knowledge, and probabilistic knowledge.

Modern Artificial Intelligence techniques currently have the ability to monitor limited aspects of training simulations. There are difficulties in handling the large temporal and dependent actions of real-time simulations. A new technique called Template Based Reasoning (TBR) has been developed to specifically assist in the interpretation portion of student actions in real-time simulations. In this approach, templates represent a group of attributes, actions or features that define a particular behavior of the student being monitored. How well a student's current behavior matches a particular template could provide a measure of confidence that the student is carrying out the procedure represented by that template. These templates are used to track student actions as they relate to the training goals. The student would progress through templates much as they would progress through scenarios in lessons. This direct student monitoring and evaluation can provide real-time feedback that is available for the instructor. By presenting the current template status, an instructor may view the student's progress and performance through the lessons. This is the focus of research at the University of Central Florida to assist instructor efficiency by offloading some of the workload of the instructor. This permits the instructor to concentrate on other important areas of simulation.