The rise of big data and predictive analytics continues to proliferate throughout many industries across the globe and the commercial insurance industry is no exception. One of the core operations of any insurance company is underwriting, where underwriters make important decisions about which risks the insurance company accepts (or rejects), at what price and what terms. These decisions have traditionally been made through a balance of human intuition, judgment and actuarial science but now, with the ability to capture large bodies of data and the application of sophisticated algorithms, insurance companies are developing new decision-making tools that may change the traditional underwriting process and methodology. This exploratory study examines the cognitive and emotional reactions of underwriters to these new tools. Although a limited amount of research exists dealing with human reactions to computer-based decision making, a near complete gap exists examining the interaction of commercial casualty underwriters and predictive analytics. A quantitative research methodology was used as the framework for an online survey was completed by 46 commercial casualty underwriters from various insurance companies. Purposeful sampling was used to select participants for the study and a simple statistical analysis was used to develop inferences about the population. The prominent finding showed that underwriters often acquiesced to the output produced by algorithmic based tools when they did not agree with the result. This in turn caused frustration and ambivalence toward the model. In addition, the majority of underwriters did not always agree with the model setting the stage for more frustration. Interestingly the data showed a notable split in how underwriters do their job now in the face of predictive analytics. The majority are still underwriting the same and considered predictive analytics as a complementary piece, rather than an exclusive one, signaling that underwriters still had meaningful influence in the decision making process while a sizable portion spent less time underwriting knowing the model would ultimately trump their decision.