NLP-powered feedback targeting calls for less extensive hand-tuning requirements for tutoring software, which allows for a wider variety of curricula as well as a wider reach to more topics, with the same amount of time spent developing the software. Additionally, while a single multiple-choice response was not adequate for providing effective feedback, a single sentence response proved to be enough information to select effective feedback. This allows for more frequent and relevant feedback. Furthermore, NLP-powered feedback targeting has the advantage of generality, unlike multiple-choice-powered feedback selection. This opens many doors for making curricula more flexible and reducing the amount of labor required for developing tutoring software. In the future, hopefully, these techniques can be applied to more complex situations. This method could prove helpful in several areas including physics and medicine. Although the NLP technique used in this study promoted substantial learning gains, it is important to test this model in more complex situations. Future work will hopefully incorporate NLP feedback targeting mechanisms into existing educational software systems, allowing for higher learning gains and better feedback for students.