We examined how the minority’s perceived (i.e., not real) expertise affects group discussion and performance. In two experiments, participants were randomly assigned to interacting groups in which the minority faction was perceived as either expert or not. Groups performed a decision task that involved solving a murder mystery. Both experiments showed that minorities perceived as expert (vs. not perceived as expert) made majority individuals acquire more accurate private judgments after group discussion, although the public group decision was not more accurate. In parallel, perceived expertise made minority members change their own judgments less. Experiment 1 also showed that minorities’ questioning behaviors mediated the effect of minorities’ perceived expertise on majority members’ private accuracy. Experiment 2 further showed that majority members’ deeper processing was also a mediator. Thus, minorities with perceived expertise serve as a catalyst, increasing the quality of majority members’ cognitions, but not their own.