Cracking the AI Code: Understanding the Boundaries of AI-Driven Promotions in Consumer-Packaged Goods

Abstract

Artificial intelligence (AI) holds great promise for improving marketing decision-making, yet its real-world performance implications remain underexplored. This study investigates how AI-informed promotional planning tools affects sales outcomes in the consumer-packaged goods (CPG) sector using a large-scale dataset provided by a leading European firm. Drawing on over 65,000 SKU-level observations across three national markets and several product categories, we examine how varying levels of AI implementation influence sales performance and identify boundary conditions determining its effectiveness. Our findings reveal significant variation in AI’s performance effects: in some categories, AI adoption exhibits U-shaped or inverted U-shaped relationships with sales, while in others, results remain consistently neutral or even negative. Moreover, market leadership status strengthens the positive effects of AI, underscoring the importance of organizational context. By integrating an empirics-first approach with the dynamic capabilities perspective, this study advances understanding of how AI-enabled decision tools influence business outcomes in digitally transforming ecosystems.