Paper Type
Complete
Paper Number
PACIS2026-2004
Description
Despite enterprises continuing to invest heavily in AI, many initiatives fail to scale or generate sustained business value. Terming this the AI-investment paradox, we argue that it persists because firms govern AI as a broad technology program rather than as a set of discrete, investable decision opportunities embedded within workflows. We address this issue by developing a decision-centric portfolio framework for governing enterprise AI investments. Our framework introduces AI-Investable Process Nodes (AIPNs) as bounded decision points where AI can alter expected outcomes and where benefits, risks, and costs can be assessed ex ante. We formalize node-level value through Expected Net Benefit, then show how AIPNs can be staged using real options logic and assembled into a broader portfolio through risk-return principles. In doing so, our framework offers a pathway to resolving the AI-investment paradox by linking AI investments more explicitly to identifiable, governable, and accumulative sources of business value.
Recommended Citation
Mathur, Abhinav; Kathuria, Abhishek; and Chaturvedi, Devina, "Governing Enterprise AI Investments: A Decision-Centric Portfolio Framework" (2026). PACIS 2026 Proceedings. 8.
https://aisel.aisnet.org/pacis2026/practioner/practioner/8
Governing Enterprise AI Investments: A Decision-Centric Portfolio Framework
Despite enterprises continuing to invest heavily in AI, many initiatives fail to scale or generate sustained business value. Terming this the AI-investment paradox, we argue that it persists because firms govern AI as a broad technology program rather than as a set of discrete, investable decision opportunities embedded within workflows. We address this issue by developing a decision-centric portfolio framework for governing enterprise AI investments. Our framework introduces AI-Investable Process Nodes (AIPNs) as bounded decision points where AI can alter expected outcomes and where benefits, risks, and costs can be assessed ex ante. We formalize node-level value through Expected Net Benefit, then show how AIPNs can be staged using real options logic and assembled into a broader portfolio through risk-return principles. In doing so, our framework offers a pathway to resolving the AI-investment paradox by linking AI investments more explicitly to identifiable, governable, and accumulative sources of business value.
Comments
16-Practitioner