AI-Powered Digital Car Showrooms: An Implementation of Intelligent Systems Transforming the Automotive Retail Experience

This paper described AutoVerse, a full-stack intelligent vehicle marketplace that demonstrates how carefully targeted AI integration can produce meaningful improvements to the user experience in a domain long characterized by static, analytically passive interfaces. Through the coordinated embedding of a valuation engine, a similarity-driven recommendation module, and a transfer-learning-based condition classifier within a unified, modular platform, AutoVerse directly addresses the four core deficiencies identified in existing solutions: pricing opacity, discovery friction, condition ambiguity, and fragmented tooling. The platform’s technology stack—FastAPI, Streamlit, PostgreSQL, and TensorFlow—consists entirely of open-source, production-grade components, demonstrating that sophisticated intelligent applications need not depend on proprietary infrastructure to achieve real-world quality. The modular three-tier organization provides a foundation for independent scaling and iteration, supporting long-term platform evolution without architectural overhaul. Empirical results confirm that each AI subsystem performs effectively within its designated role and that their co-presence within a single interface yields a user experience qualitatively superior to what any component could provide operating in isolation.

[1] F. Ricci, L. Rokach, and B. Shapira, Recommender Systems Handbook, 2nd ed. New York, NY, USA: Springer, 2015. [2] M. J. Pazzani and D. Billsus, “Content-based recommendation systems,” in The Adaptive Web, P. Brusilovsky, A. Kobsa, and W. Nejdl, Eds. Berlin, Germany: Springer, 2007, pp. 325–341. [3] A. Singhal, “Modern information retrieval: A brief overview,” IEEE Data Eng. Bull., vol. 24, no. 4, pp. 35–43, 2001. [4] I. Pudaruth, “Predicting the price of used cars using machine learning techniques,” Int. J. Inf. Comput. Technol., vol. 4, no. 7, pp. 753–764, 2014. [5] T. Chen and C. Guestrin, “XGBoost: A scalable tree boosting system,” in Proc. 22nd ACM SIGKDD Int. Conf. Knowl. Discovery Data Mining, San Francisco, CA, USA, 2016, pp. 785–794. [6] A. Krizhevsky, I. Sutskever, and G. E. Hinton, “ImageNet classification with deep convolutional neural networks,” Commun. ACM, vol. 60, no. 6, pp. 84–90, Jun. 2017. [7] M. Sandler, A. Howard, M. Zhu, A. Zhmoginov, and L.-C. Chen, “MobileNetV2: Inverted residuals and linear bottlenecks,” in Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR), Salt Lake City, UT, USA, 2018, pp. 4510–4520.