Abstract: Traditional car management system models are field-based, with complex interfaces that necessitate extensive training to master. The proposed project aims to design an intelligent car management system that integrates agentic artificial intelligence via Google's Gemini 2.0 Flash approach, FastAPI technology, and SQLite database services. Thus, this proposed model will allow natural language processing to undertake full CRUD tasks to implement commands such as "show available cars" autonomous execution as database queries without human interference Performance analysis indicates a data-entry time reduction of up to 60%, response times of less than a second, and accuracy of 95% with respect to interpretation of user inputs, sufficient to handle up to 50 concurrent users. The underlying technology, Agentic, relies on pattern recognition software that uses FC calls to manage carry-over effects in interpretation guaranteeing data integrity.

Keywords: Agentic AI, Natural language processing, Fleet Management, Conversational AI, Database Automation, Function-Calling

Downloads: | DOI: 10.17148/IARJSET.2025.121243