The Business Value of AI: Decision Productivity, Coordination Costs, and the Governance of Intelligent Systems

The paper argues that AI's uneven firm-level productivity record echoes the classic "IT productivity paradox," and explains it with a formal model: AI creates value only when it improves decision productivity (impact × velocity × scale) by more than it adds organizational friction (coordination complexity and governance overhead). When decision gains outpace these costs, firms achieve structural advantage; when they don't, AI remains cosmetic, becomes bureaucratic, or actively destroys value.
Its core contributions:
Reframes AI as an organizational design intervention, not a technology purchase.
Positions coordination and governance costs as the moderating variables that determine whether AI pays off.
Derives testable hypotheses connecting decision redesign to measurable firm-level impact.
In one sentence: AI's economic payoff depends less on the technology itself than on whether redesigned decisions beat the coordination and governance costs the technology introduces.