Paper Type
Short
Paper Number
PACIS2026-1282
Description
AI tools are increasingly used to support IS portfolio governance through automated evaluation and prioritization, yet portfolio decisions involve structural interdependencies that AI systems struggle to encode. This research-in-progress examines how enterprise IS portfolio managers navigate tensions between AI automation and human judgment under varying contextual conditions. Drawing on paradox theory, the study employs experimental vignettes with 32 portfolio professionals (192 responses) and applies configurational analysis to identify how combinations of technical (AI explainability), situational (decision urgency, task complexity), and cognitive (AI-human congruence) conditions shape paradox navigation strategies. Preliminary findings suggest context-specific divergence from established algorithmic decision-making patterns: AI-human disagreement associates with integration rather than aversion, and task complexity associates with synthesis rather than automation reliance. These patterns are consistent with decision object interdependence as a promising boundary condition for algorithmic decision-making theory, with implications for AI system design in portfolio governance contexts.
Recommended Citation
Puthenpurackal Chakko, Joseph, "Exploring Paradox Navigation Configurations in AI-Augmented Portfolio Decision-Making" (2026). PACIS 2026 Proceedings. 4.
https://aisel.aisnet.org/pacis2026/it_strategy/it_strategy/4
Exploring Paradox Navigation Configurations in AI-Augmented Portfolio Decision-Making
AI tools are increasingly used to support IS portfolio governance through automated evaluation and prioritization, yet portfolio decisions involve structural interdependencies that AI systems struggle to encode. This research-in-progress examines how enterprise IS portfolio managers navigate tensions between AI automation and human judgment under varying contextual conditions. Drawing on paradox theory, the study employs experimental vignettes with 32 portfolio professionals (192 responses) and applies configurational analysis to identify how combinations of technical (AI explainability), situational (decision urgency, task complexity), and cognitive (AI-human congruence) conditions shape paradox navigation strategies. Preliminary findings suggest context-specific divergence from established algorithmic decision-making patterns: AI-human disagreement associates with integration rather than aversion, and task complexity associates with synthesis rather than automation reliance. These patterns are consistent with decision object interdependence as a promising boundary condition for algorithmic decision-making theory, with implications for AI system design in portfolio governance contexts.
Comments
11-Strategy