AI-Driven Optimization of Debt Collection Efficiency in Accounts Receivable: An Action Research Case Study of DK Company

Abstract

This study takes DK Company, a Chinese SME manufacturer, as a case and adopts the action research method to explore the practical application of artificial intelligence (AI) in optimizing the debt collection process in accounts receivable management. DK Company has long struggled with delayed customer payments and a days-sales-outstanding (DSO) ratio well above the industry median. In collaboration with the research team, DK developed and deployed an AI-driven collection system that segments debtors by default risk, predicts payment behavior, and generates personalized, channel-specific dunning strategies. Through continuous interaction between researchers and the company’s finance team and multiple rounds of iterations, this system has significantly enhanced the collection efficiency and effectively improved the cash flow situation. The empirical results demonstrate substantial improvements across key operational indicators: DSO decreased by 25.7%, bad debt ratio declined by 41%, and an additional RMB 1.5 million in cash flow was recovered within six months. Beyond quantitative gains, the intervention fostered user acceptance and organizational learning, with cross-departmental collaboration and data-driven decision-making capabilities notably enhanced. This study theoretically expands the application boundaries of artificial intelligence in the field of financial management and provides an operational reference for the accounts receivable management in similar manufacturing settings at the practical level. The main limitation of the research lies in its reliance on a single case. In the future, the universality of the conclusion can be further verified through multi-case comparisons or cross-industry verifications. This study was motivated by the persistent cash flow constraints experienced by small and medium-sized manufacturing enterprises in China. In many cases, reliance on manual debt collection procedures led to operational inefficiencies and intensified financial pressure. The primary contribution of this research lay in the development and empirical examination of a practical AI-driven framework. The findings demonstrated that the proposed framework improved key operational performance indicators while facilitating organizational adaptation, thereby providing a replicable reference model for enterprises operating in comparable contexts.