Assessing car damage is a crucial task in the automotive and insurance sectors, significantly impacting automation in claims processing, repair workflows, and overall customer satisfaction. To this, various works using machine learning or deep learning have been implemented to detect the severity level or location of damage; however, an integrated approach providing damage status, location, and severity level from a single input image remains absent. This work introduces a hybrid deep learning-based approach leveraging transfer learning and fine-tuning techniques that integrates three specialized Deep Learning models to accurately evaluate damage status, location, and severity from a single image input. The dataset used for experimentation consists of 1650 labeled images of diverse qualities and complex patterns of damaged and undamaged cars. The results highlight an average accuracy of 79.38%, outperforming SOTA using advanced deep learning architectures for fully automated damage detection and classification.