Expeditiously moving new Air National Guard (ANG) recruits through the initial training pipeline is critical to generating readiness and to the optimal use of personnel resources. In executing its training mission, the ANG faces challenges in accurately forecasting demand for basic military training, technical school slots and officer training school. Currently, the ANG collects data through various methods to establish future fiscal year requirements for tech training school seats. The ANG is experiencing significant wait times (more than a year) for critical career fields due to a lack of unit participation in the current data call process or gaps in manning. This uncertainty leads to delays in the deployability of new accessions, degraded readiness, and inefficient resource expenditure while new members await training. The ANG desires to build upon a more effective toolkit for predicting demand for scarce training resources to improve its readiness. The Institute for Defense Analyses (IDA) uses the Retention Prediction Model, a machine learning capability that IDA developed, to produce these forecasts.