Answers are grounded in this paper and cite their sources.
Digital platforms generate customer information signals, yet organizations struggle to leverage them for transformation. Drawing on information processing theory and absorptive capacity theory, we examine how customer information signals differentiated by tangible versus intangible attributes shape external knowledge search from proximate ecosystem partners and distant institutional sources, how digital capacity enables that search, and under what resource conditions acquired knowledge contributes to digital transformation. We test our model combining survey data from 149 organizations with generative AI aspect-based sentiment analysis of 24,913 customer reviews validated across nine large language models. Tangible attribute signals selectively trigger search toward distant institutional sources while intangible signals reduce it, demonstrating content-contingent attention allocation where signal domain determines search direction. Digital capacity operates as potential absorptive capacity enabling both proximate and distant search, with stronger effects on proximate sources. Proximate partner knowledge positively affects digital transformation, an effect that strengthens after endogeneity correction, while distant institutional knowledge shows no direct transformation effect once endogeneity is addressed. Non-financial resources strengthen the proximate knowledge to transformation relationship and financial resources strengthen the distant knowledge to transformation relationship, revealing dual absorptive capacity pathways through exploitation and exploration respectively. This study advances information processing theory by demonstrating that signal content type determines search direction and extends absorptive capacity theory by identifying how digital capacity enables knowledge acquisition and how resource complementarity varies by knowledge source cognitive distance. Methodologically, we validate generative AI text analytics achieving 96.8 % accuracy with open-weight alternatives providing statistically equivalent performance for reproducibility.
No takes yet. Share an insight, caveat, or question.
Codesso et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: