We used a newly developed experimental paradigm to investigate aftereffects of completed intentions on subsequent performance that required the maintenance and execution of new intentions. Participants performed an ongoing number categorization task and an additional prospective memory (PM) task, which required them to respond to PM cues that differed from standard stimuli in 1 particular visual feature. Although the feature defining the to-be-acted-upon PM cue changed in each block, the irrelevant PM cue of the previous PM task block was occasionally repeated in the subsequent block. In 4 experiments we found that performance in the ongoing task was substantially slowed for repeated PM cue trials compared to oddball trials, which also differed in a visual feature from standard stimuli but never served as PM cues. This aftereffect decreased as a function of delay after intention completion. These findings indicate that intentions can exhibit persisting activation even after they have been completed and may interfere with the execution of the new relevant task. Possible mechanisms and boundary conditions of this intention deactivation failure are discussed.