Customer segmentation is essential for businesses to develop targeted strategies based on purchasing behavior. Traditional clustering methods, such as K-Means, have been widely applied in this area; however, they struggle with high-dimensional data and complex, non-linear customer relationships. To address these limitations, this paper explores the integration of deep learning techniques—specifically, autoencoders for dimensionality reduction—with conventional clustering algorithms. Autoencoders compress high-dimensional data into a more manageable format while preserving critical patterns, enabling more accurate and meaningful segmentation. This hybrid approach is applied to a large retail dataset, demonstrating its effectiveness in uncovering distinct customer groups with greater precision. The resulting segments offer deeper insights into customer preferences and behaviors, allowing businesses to refine their marketing campaigns, personalize customer experiences, and optimize product offerings. By leveraging deep learning alongside traditional clustering, companies can enhance decision-making and drive customer-centric growth in an increasingly data-driven market.