This paper addresses a central question in strategy and entrepreneurship: how does artificial intelligence (AI) create value for organizations? Prior research emphasizes demand-side mechanisms, most notably data network effects, whereby user interactions generate data that improve algorithmic prediction and reinforce adoption. The emergence of AI agents capable of autonomously performing goal-directed tasks introduces a qualitatively different mechanism. We propose the intelligence loop, a process wherein agentic capability, comprising self-learning, self-adapting, and self-improving, drives continuous intelligence improvement for autonomous task execution. We define intelligence as the demonstrated competence in perceiving, reasoning, and acting to achieve goals across relevant environments, and introduce minimum viable intelligence as the threshold at which the intelligence loop activates. The loop may decay, self-sustain, or self-accelerate. Entrepreneurial opportunity derives not only from unmet customer demand but also from supply expansion, i.e., deployment of autonomous agents to address limited supply of human capital. The framework offers theoretical insights into AI-driven value creation and has implications for entrepreneurship and strategy in the age of AI agents.

ggrace2@ceibs.edu