Developing Entrepreneurial Capabilities with Large Language Models: A Mixed-Methods Analysis of Microenterprise Training in Barranquilla's Business Ecosystem

Abstract

This research examines the impact of Large Language Models (LLMs) on entrepreneurial development among microentrepreneurs in Barranquilla, Colombia. Through a mixed-methods investigation, the study evaluates two LLM interventions, Claude Sonnet and Chat GPT 3.5, in comparison to conventional entrepreneurship education approaches. A cohort of 105 microentrepreneurs engaged in a structured training program, with outcomes assessed across multiple entrepreneurial competency domains. Quantitative analysis utilizing Hierarchical Linear Modeling and Bayesian Analysis revealed significant enhancements in business model development, market analysis capabilities, and strategic decision-making among LLM intervention participants. Mediation analysis identified increased motivation and self-efficacy as key mechanisms driving these improvements. Qualitative insights from semi-structured interviews highlighted participants' improved ability to identify market opportunities and develop innovative solutions to business challenges. The research contributes to our understanding of AI-enhanced entrepreneurship education, particularly in emerging market contexts. These findings offer valuable guidance for policymakers and educators seeking to leverage digital technologies to support microenterprise development in resource-constrained environments.

Received: 4 January 2024 / Accepted: 22 November 2024 / Published: 05 January 2025