Which technical and financial factors most heavily influence managers' decisions to invest in artificial intelligence?

Host: Welcome to A.I.S. Insights — powered by Living Knowledge. I'm Anna Ivy Summers. Today, we are exploring a study titled "Valuing the Uncertain: Identifying AI Investment Preferences in Organizations" by Victoria Sofia Santos Gabriel, presented at the International Conference on Information Systems in 2026. This study investigates how organizational leaders evaluate and trade off project attributes when deciding whether to allocate capital to artificial intelligence versus traditional business intelligence. Alex, welcome.

Expert: Thanks, Anna. It is great to be here.

Host: To start off, what is the core problem leaders face when deciding to invest in AI?

Expert: Traditionally, IT investments rely on established financial appraisal frameworks, like return on investment or net present value. But artificial intelligence fundamentally shifts that logic. AI systems are probabilistic, constantly evolving, and require extensive complementary organizational adaptations to deliver value. Because outcomes and integration costs are hard to measure upfront, managers face significant uncertainty when choosing which projects to fund.

Host: How did the researcher examine these decision patterns?

Expert: The study employed a discrete choice experiment with a split-sample design. A sample of sixty-three experienced professionals evaluated four hundred and forty-one decision scenarios. Participants were randomly assigned to evaluate either AI or Business Intelligence projects. Each scenario compared two options with varying levels across six attributes: implementation costs, probability of success, expected impact, system autonomy, modularity, and adaptability.

Host: And what were the key findings from this experiment?

Expert: The findings revealed a clear divergence in evaluation logic between AI and BI. For traditional BI, managers followed classic risk-averse economic behavior: higher probability of success was by far the strongest driver, higher costs significantly reduced project selection, and higher operational autonomy was preferred.

Host: Did AI follow that same risk-averse pattern?

Expert: Not at all. For AI projects, technical architecture dominated the decision. High modularity and high adaptability were statistically significant positive drivers of choice. In fact, high modularity increased selection odds nearly fourfold. Interestingly, autonomy had no statistically significant effect on AI choices, which may reflect the tension between AI efficiency and managerial loss of control.

Host: What about costs and the expected strategic impact of AI?

Expert: Expected impact—whether framed as process improvement, decision support, or competitive advantage—did not reach statistical significance in predicting AI choices. Furthermore, implementation costs showed a small, statistically significant positive association with AI selection. Rather than acting as a deterrent, higher costs in AI might signal to decision-makers that the project includes the necessary budget for broader organizational transformation and readiness.

Host: What about the willingness-to-trade analysis? Did managers tolerate lower success rates for high-upside projects?

Expert: The study hypothesized that managers would accept lower success odds or higher costs for high strategic impact, but the interaction models did not statistically support those hypotheses. Instead, the willingness-to-trade analysis showed that managers were willing to absorb substantial cost increases specifically for high modularity and adaptability. Rather than chasing vague future rewards, leaders appear to use modular and adaptable architectures as a hedge to retain flexibility and manage uncertainty.

Host: What does this mean in practice for technology leaders and business professionals?

Expert: There are two clear practical takeaways. First, when pitching or assessing an AI initiative, focus heavily on architectural robustness. Emphasizing adaptability and modular design addresses managerial concerns about lock-in and risk under conditions of technological uncertainty. Second, do not assume traditional cost-minimization frameworks apply identically to AI. Framing costs in terms of organizational capabilities, process redesign, and systemic adaptability aligns much closer with how capital allocators assess AI initiatives.

Host: That provides useful clarity on how digital investment choices shift with emerging technologies. Thank you for breaking down the study, Alex, and thank you to our listeners for tuning in to A.I.S. Insights — powered by Living Knowledge. Join us next time.