The question of how people make use of automation to support their decision making is becoming increasingly important. As computers provide ever greater input to the collection, analysis and interpretation of data, so they are more likely to be partners in decision making. However, when automation makes recommendations that the human disagrees with or that might be based on erroneous analysis, then this could result in a change in decision strategy. It is not simply a matter of ignoring or rejecting the recommendation but rather a matter of deciding how best to make use of the automation's output. By modeling information search and decision strategies under different levels of information reliability, we demonstrate that it makes sense to adapt decision strategy to the information context.