Publication Details
Abstract
The diffusion of artificial intelligence (AI) into marketing functions has reshaped the way firms allocate budgets, target audiences and interact with customers in the digital economy. Yet empirical evidence quantifying how much AI-based marketing improves economic efficiency, and through which behavioural channels it operates, remains fragmented. This study develops and tests an integrated econometric framework linking AI marketing capabilities to firm-level economic efficiency and to individual-level consumer behaviour. Two complementary datasets were analysed: a firm-level survey of 214 enterprises that have adopted AI-enabled marketing tools, and a consumer-level survey of 386 digital buyers. Ordinary least squares regression with heteroskedasticity-robust standard errors was applied to both samples, supported by reliability and validity diagnostics, multicollinearity checks and robustness tests. The firm-level model explains 67.4% of the variance in the return on marketing investment (ROMI); AI adoption intensity is the strongest predictor (β = 0.412, p < 0.001), followed by predictive analytics and data quality. The consumer-level model explains 59.3% of the variance in purchase intention, with perceived personalization (β = 0.342) and recommendation relevance (β = 0.291) emerging as the dominant drivers, while privacy concern exerts a significant negative effect (β = −0.174). The findings indicate that the economic returns of AI marketing are realised primarily through improved targeting precision and personalization rather than through automation alone, and that trust and privacy management are necessary boundary conditions. The paper contributes a measurable, replicable specification for evaluating AI marketing effectiveness and offers managerial and policy implications for emerging digital economies.