Brain tumors are among the most critical and life-threatening neurological conditions, affecting individuals across all age groups. According to recent global cancer reports, over 308,000 new cases of brain and central nervous system tumors are diagnosed annually, underscoring the urgent need for early and accurate detection. Although MRI and CT scans remain the gold standard for diagnosis, manual interpretation is time-consuming and depends heavily on radiologists' expertise, resulting in subjective variability and delayed clinical decision-making. This study introduces a comprehensive AI-powered framework that integrates preprocessing, data augmentation, CNN-based tumor segmentation, and transfer-learning-based classification to automate brain tumor detection. Using U-Net for segmentation and ResNet/DenseNet architectures for classification, the proposed system demonstrates strong performance with a Dice score of 0.91, an IoU of 0.86, a classification accuracy of 96.3%, and an average F1-score of 0.95 across glioma, meningioma, and pituitary tumor types. Evaluation algorithms, deployment architecture, and workflow diagrams are provided to ensure methodological transparency. The findings confirm the significant role of AI in increasing diagnostic accuracy, reducing interpretation time, and supporting clinical decision-making. The framework offers a reliable, efficient, and scalable solution for real-world medical imaging applications.