Refurbished Goods Shopping Agent

Abstract: The increasing demand for refurbished electronic products has created challenges in product discovery, pricing transparency, and efficient management due to variations in condition and availability. Traditional e-commerce systems depend on manual search and filtering, which often leads to increased user effort and reduced clarity. This paper proposes Second Life, an agentic AI-based refurbished goods shopping system that enables conversational interaction for both product exploration and management. The system employs a language model to interpret user intent and invokes validated backend tools to perform database operations such as retrieving product details, checking availability, applying discounts, and updating or removing listings. A structured service-repository architecture is adopted to ensure secure execution and data consistency. The developed system demonstrates improved interaction efficiency, accurate data handling, and reduced manual intervention. The results indicate that agentic AI can effectively support refurbished e-commerce platforms by enhancing usability and automating operational workflows.

Keywords: Agentic AI, Refurbished Electronics, Conversational AI, Tool-Based Execution, ECommerce Automation, Database Management.