The wisdom of the crowd is the idea that aggregated group decisions can outperform most or even all of the individuals in the group. We argue that cognitive models, built on an understanding of people's judgment and decision making, can further improve the wisdom of the crowd in four ways. The first way is that they can infer and upweight expertise among individuals. The second is that they can be used to debias cognitive processes by inferring what people know from how they behave. The third is that they can provide a representational scaffolding for combining knowledge that is multidimensional in nature and distributed across individuals. Finally, cognitive models can maintain the diversity of a crowd by producing predictions that act as surrogates for unavailable behavioral data. We demonstrate these ideas in a range of decision-making settings including probability estimation, ranking, spatial knowledge, competitive bidding, and sequential decision making. We also highlight how studying these applied knowledge aggregation problems helps identify new creative directions in the development of basic theories and models of decision making.