In two experiments, we explored whether the retrieval processes underlying event-based prospective memory can be distinguished from those underlying vigilance. Participants performed an ongoing task (either a lexical decision task, Experiment 1 and 3, or a categorization task, Experiment 2) while at the same time they had to remember to stop performing the ongoing task whenever a particular target-stimulus appeared on the computer screen (background task). There were two target stimuli, each appearing 4 times across the ongoing task. Instructions and training induced participants to encode the background task either as a prospective memory task or as a vigilance task. Results revealed important processing differences between prospective memory and vigilance processes. The time to respond to the ongoing task was systematically slower in the vigilance than in the prospective memory conditions. However, prospective memory conditions did not differ from control (ongoing task only). Accuracy in the background task complemented RT data in the ongoing task in that more errors were observed in the prospective condition as compared to the vigilance condition. The differences were not due to a speed-accuracy trade-off between the ongoing and the background tasks, nor to the differences in training. Most important, repetition priming was observed across the four target presentations in the prospective memory condition but not in the vigilance condition. The results are consistent with the hypothesis that eventbased prospective memory and vigilance processes differ as to the degree of conscious monitoring that they require, with prospective memory being based more on automatic retrieval of the cue-action association and vigilance being based more on active search for the target. Key words: prospective memory, vigilance processes, automaticity.