Orchestrating Tradition: How Family Firms Build Grounded AI Capabilities

As Artificial Intelligence (AI) is increasingly portrayed as a general-purpose technology, the strategic issue shifts from adoption to appropriation. This paper addresses this appropriation puzzle in family firms, exploring how they can prevent identity dilution while leveraging AI. We develop a process model explaining how family firms orchestrate tradition and socioemotional wealth (SEW) resources to transform generic AI into firm-specific, appropriable capabilities. Through three sequential phases—Structuring, Bundling, and Leveraging—mediated by Interiorizing (identity-aligned filtering) and Reinterpreting (generative actuation), family firms build a Grounded AI Capability. This sociotechnical capability integrates tradition-embedded knowledge with AI technologies through supervised fine-tuning and reinforcement learning. The model advances resource orchestration research by revealing phase transition mechanisms, extends the SEW as a strategic resource perspective, and recontextualizes innovation-through-tradition for the AI era.