How Employees Use AI to Explore and Confirm: A Study of Cognitive Drivers, Hybrid Collaboration, and Performance

Research on artificial intelligence (AI) has traditionally explored its adoption, usage and acceptance by employees; however, there is limited knowledge regarding the cognitive processes employed when making decisions with AI. Exploratory and Confirmatory uses of AI are differentiated in this study. The former includes generating alternative options and increasing a decision-makers’ decision space. The latter is used to verify and/or refine existing judgments. Analytical decision-making, uncertainty avoidance, creativity and mindfulness were all considered cognitive orientations which could potentially affect an employee’s mode of using AI. Also, the impacts of each mode on both decision satisfaction and performance will be investigated. Data were collected via a cross-sectional online survey of employees who used AI in decision-making. A total of 1472 surveys were returned (mean age 30.5, mean tenure 9.5 years, 52% female). Participants rated multiple-item scales on exploratory AI use (2 items; α=.77), confirmatory AI use (2 items; α=.76), creative decision orientation (3 items; α=.87), mindfulness (9 items; α=.81), analytical decision orientation (9 items; α=.86), uncertainty avoidance (5 items; α=.90), decision satisfaction (4 items; α=.86), task performance (4 items; α=.87), and hybrid human-AI decision making (1 item). Composite scores and reliabilities were acceptable, with measures showing adequate convergent validity (AVE > .50). Data analysis used linear regression with heteroskedasticity-consistent (HC3) robust errors and PROCESS macro bootstrapping (5,000 resamples). Hypotheses were tested with direct-effects models controlling for age, sex, tenure, and education. PROCESS Model 4 assessed parallel mediation, while Model 6 evaluated serial mediation through hybrid decision-making, decision satisfaction, and performance. The findings indicated that creative decision orientation and mindfulness are strong predictors of exploratory AI use, which in turn is a strong predictor of hybrid human-AI decision-making. Additionally, hybrid decision-making and confirmatory AI use each contributed to greater levels of decision satisfaction. Decision satisfaction was identified as a strong predictor of task performance. Both exploratory and confirmatory AI use indirectly affected task performance through satisfaction. The primary pathway from exploratory AI use through hybrid decision-making to satisfaction was confirmed. Analytical decision orientation and uncertainty avoidance were not found to significantly predict confirmatory AI use in the combined model. Thus, it appears possible that additional situational or design variables may be more important. Overall, this study indicates that differentiating how employees are using AI - to explore new possibilities or to confirm existing judgments - is crucial for understanding the cognitive mechanisms underlying effective AI-aided decision-making. The results advance human-AI collaboration theory by differentiating exploration- and confirmation-oriented AI use. For researchers, it offers a cognitive framework for understanding how employees adopt AI in decision-making, beyond just automation or augmentation.