Knowledge over technology? Unpacking the nonlinear effects of absorptive capacity on business model innovation in Liberian SMEs

While digital transformation is often heralded as the primary driver of innovation, this study argues that in resource-constrained environments, the ability to integrate and utilize knowledge is a more critical yet fundamentally nonlinear determinant of Business Model Innovation (BMI). Drawing on survey data from 300 manufacturing SMEs in Liberia, we employ Artificial Neural Networks (ANN) to move beyond traditional linear assumptions and uncover the complex predictive relationships between firm capabilities and BMI. The findings reveal a stark capability hierarchy: absorptive capacity emerges as the dominant predictor of BMI, with an average importance of 47% across all dimensions, peaking at 69.7% for value creation. In contrast, the importance of digital capability is highly volatile ranging from 3% in value creation to 44.9% in value delivery confirming its context-dependent role as an enabler rather than a driver. Leadership capability provides a stable yet secondary foundation, averaging 29.4% with peak influence in value capture. This study makes a key intellectual contribution by reframing innovation in developing economies as a primarily knowledge-driven, nonlinear process: while digital tools enable innovation, it is a firm's knowledge integration capacity that ultimately determines the intensity and success of business model transformation. These findings shift the policy and managerial focus from technology adoption per se toward the cultivation of deeper organizational learning mechanisms particularly critical in post-conflict contexts where institutional voids constrain traditional innovation pathways.