Leveraging Large Language Models for Compliance and Productivity: Economic Implications of AI Adoption in the U.S. Small Business Sector

Abstract

This paper investigates the economic implications of adopting Large Language Models (LLMs) to automate compliance and administrative functions within the U.S. small business sector. Small and medium-sized enterprises (SMEs) currently face disproportionate regulatory burdens that consume approximately 19 percent of operating budgets and cost over $50,000 per employee annually. By transforming compliance from a labor-intensive cost center into a scalable digital process, LLM-enabled automation offers a structural remedy to the SME productivity gap. The analysis estimates that widespread adoption could reallocate $50 to $90 billion annually from non-productive administrative tasks to high-value productive uses. Beyond immediate cost savings, the paper demonstrates that this technological shift promotes capital deepening, as firms redirect resources from recurring operational expenses toward long-term intangible capital formation. The study concludes that supported by adaptive federal policies and deregulation, AI-driven compliance automation will serve as critical economic infrastructure, enhancing total factor productivity, fostering entrepreneurship, and strengthening U.S. competitiveness in global markets.