Imagine a synthetic environment that automatically adapts to better achieve the goals of its intended use. A training simulator that understands a trainee’s strengths and weaknesses and changes the tasks presented accordingly. The value of infusing Artificial Intelligence (AI) and Machine Learning into simulation-based training systems is undeniable. Previous articles clearly show this, for driver training systems and Live, Virtual, and Constructive (LVC) training environments, as well as for many other applications. Yet, what does it take for Artificial Intelligence and Machine Learning to reach its potential?  Rigorous, measured training performance metrics that can be compared to a standard or correct response  are necessary for a system to adapt training scenarios to each trainee. For AI to effectively adjust to a student, this assessment needs to provide significant insight into the capabilities and limitations of the trainee and provide metrics that span a spectrum (not a simple binary result, e.g., pass or fail). These metrics need to be measured during training, assessed by the AI system, and then used to dynamically modify the training venue to improve training outcomes. A process to accomplish this for any simulation-based application – beginning with describing the requirement, assessing the simulation’s role, deriving associated metrics and success criteria, measuring the metrics, and characterizing how the AI could be used to adapt the synthetic environment - is described. Then, this process is applied to two simulation bookends: simulator-based driving training and LVC-based staff officer training. These proof-of-concept examples provide a means to describe associated insights, lessons learned, and useful next steps. This paper describes and provides the beginning of the technical detail needed to implement AI augmented digital simulation-based training systems (AI-ADSTS) capable of adapting to the application’s trainee or training audience in ways that make the simulation training system even more effective.