The functional components of any tutor, be it human or computer, can be broken down by process: presenting content to the learner, assessing the performance of the learner, making an instructional strategy decision, implementing this strategy, and determining the impact. While an ideal Intelligent Tutoring System (ITS) can perform all of these actions, the construction of such a system poses its own problem: authoring these functions. This is part of the reason that ITSs are frequently built as monolithic single-topic systems, rather than modular, open-architecture frameworks. An ideal authoring tool would allow a user to create, or have automatically created, all the elements of an ITS with little interaction. While this goal is far reaching, there has been significant effort in the creation and evaluation of tools to support at least one function of tutoring.

The creation of adaptive training is likely to take longer than traditional training for the foreseeable future. This is for one simple reason: more content is required. Due to the corresponding increase in effort, each organization must decide individually whether the performance improvement is worth the cost. However, developmental costs are declining due to the utilization of user tools and automated computer tools to develop the content and function of the intelligent tutor. Research in this area focuses on streamlining the process of authoring adaptive content.

This paper discusses the issues and successes in the development of authoring tools used to generate adaptive training content and functionality. This includes automated tools, such as those for unobtrusively capturing Subject Matter Expert (SME) performance for the purpose of student assessment. This also includes human user tools, such as cognitive model authoring via SME input. The authors conclude with recommendations, based on current research, providing direction to the tasks of creating adaptive content and function.