Anniciello, Arianna (2025) On the Benefits of AI Investments: A Dynamic Multi-Criteria Framework for Strategic Evaluation. [Tesi di dottorato]

Documento PDF Arianna_Anniciello_ITEE_38_PhD_Thesis_.pdf Download (3MB)

Collection description

Despite the exponential growth of Analytic Hierarchy Process (AHP) adoption, most industrial initiatives still struggle to demonstrate sustained value. Proof of Concept (PoC) often succeed technically yet fail to scale, trapped between inflated expectations and underestimated organizational realities. The root cause lies in how value is conceived and measured: traditional cost–benefit analyses treat Artificial Intelligence (AI) as a static investment, ignoring the temporal mechanisms — adoption inertia, learning saturation, quality decay, and compliance barriers — that shape its lifecycle performance. This thesis introduces the Dynamic Benefits Model for Artificial Intelligence (DBAI) framework, a time-dependent model for assessing the lifecycle performance of AI initiatives across operational, organizational, and economic dimensions. The framework decomposes AI benefits into four dynamic components — adoption, learning, operational quality, and trust and compliance — each parameterized by readiness-adjusted factors derived through a fuzzy Multicriteria– Majority Judgment (MJ) aggregation of expert knowledge. On the cost side, DBAI distinguishes non-recurring investments (deployment and integration) from recurring expenditures (usage, obsolescence, overhead), capturing the transition from upfront effort to steady operational regimes. Our contribution lies in the explicit introduction of time dependence into both benefit and cost functions, enabling AI evaluation to move beyond static financial indicators toward a dynamic representation of how value emerges, decays, and stabilizes over time. The interplay between these dynamics reproduces the empirical J-curve of AI value realization: an initial negative phase dominated by investment, followed by recovery as adoption accelerates, learning consolidates, and governance maturity stabilizes performance. Beyond analytical precision, DBAI bridges academic rigor and enterprise practice, providing a transparent and explainable framework for portfolio-level decisions that support responsible and sustainable AI adoption.

Downloads

Actions (login required)

View Item