In a recent single-case study, we showed that a new verbal-instruction system, ensuring the automatic presentation of step instructions, was beneficial for promoting the task performance of a woman with multiple disabilities (including blindness). The present study was aimed at replicating and extending the aforementioned investigation with three adults with multiple disabilities. During Part I of the study, the new instruction system was compared with a system requiring the participants to seek instructions on their own. Two tasks were used, one per system. During Part II of the study, the new system was applied with two additional tasks. The results of Part I showed that (a) the participants had a better performance (i.e., in terms of correct steps or task execution time) on the task carried out with the new system than on the task carried out with the comparison/control system, and (b) the performance of this latter task improved rapidly when the new system was used with it. The results of Part II showed satisfactory performance with each of the two tasks carried out directly with the new system. The implications of these data were discussed. (Contains 6 figures and 1 table.)