Speech recognition can promote enhanced training procedures and reduce operating costs in training systems. For this reason, the incorporation of speech technology into training systems is becoming more prevalent. Many users of these training systems, however, are unaware of the technical capabilities of speech recognition, and therefore have unrealistic expectations which affect trainer acceptability. To prevent this, it is important for the user and developer of any training system to probe the question: "Is speech recognition appropriate for this training application?" Logicon has integrated speech technology into air traffic control training systems for nearly 15 years. In transitioning from research and development systems to fully operational trainers, experience has been gained regarding this fundamental question. This paper identifies the issues associated with determining the suitability of speech recognition for a particular training application.