Behavioral interviewing constitutes a recent innovation for improving the validity of job interviews. It involves asking so-called past-behavior questions of applicants, thus inviting them to tell a story about their behavior in a past work-related situation. Previous research in personnel selection shows that applicants often have trouble narrating complete and coherent stories. We present a procedure for extracting storytelling components (situation, task, action, results, STAR) from transcribed stories. The procedure can be used to give tailored, automatic feedback to applicants, thus potentially improving their storytelling performance.