From Startup Intention to AI-Enabled Digital Execution: Modeling How Students Turn Business Ideas into Early Computational Action

Abstract

This study examines how students convert AI driven entrepreneurial intention into early action by extending TPB/TAM with digital literacy. Here, "early action"is understood as early digital execution (EDE): whether students are already using AI-enabled computing tools to build, automate, and test parts of a business idea (e.g., generating market validation content with generative AI, automating simple workflows, assembling AI-assisted prototypes). Using a cross sectional survey of Vietnamese undergraduates and PLS-SEM, the study validates reliable constructs and strong model performance. The findings contribute to theory by integrating digital capability into TPB in the AI linked entrepreneurship context, while offering practical implications for universities to prioritize digital skills training, design learning and doing experiences with AI, and activate social and incubation networks to translate intention into action. Beyond theory, the paper proposes concrete computing design rules, pre-configured AI sandboxes, one-click automation starter kits, and short feedback-driven build sprints that operationalize the progression from Digital Literacy (DL), to Behavioral Intention (BI), and ultimately to Early Digital Execution (EA/EDE) as measurable digital execution rather than just aspiration. At a computational level, the study formalizes this mechanism as a structural model BI = f(PU, SN, DL) and EDE = f(BI), and estimates the corresponding path coefficients (β) using PLS-SEM with bootstrapped resampling. The main limitations concern the cross sectional design and sample scope; subsequent research should broaden the context and employ longitudinal tracking.