Nuclear power plants are required to ensure that conditions adverse to quality are promptly identified and corrected. For this reason, nuclear plants have corrective action programs (CAPs) to manage the identification, screening, and fixing of adverse conditions. The scope, complexity and amount of information involved in CAPs require the use of software that provides a centralized system for collecting information on issues of all levels of significance. Issues may range from minor concerns (e.g., a sidewalk crack) to critical situations that may present potentially unsafe operating conditions in the plant. Entries into the CAP are routed to relevant administrators and specialists to triage, remedy, and resolve the issues that have been raised in CAP reports. At a typical U.S. nuclear plant, the CAP may receive 10,000 or more new entries annually. These are screened by a multi-disciplinary team of senior plant personnel, requiring many labor hours each week. This review of CAP entries may vary depending on the individuals reviewing the CAP entry, their experience level and familiarity with the issues that are described. To improve the accuracy, consistency and efficiency of this process, Xcel Nuclear developed machine-learning (ML) models to identify, screen and remediate adverse conditions described in CAP entries.