To improve consumer satisfaction and boost sales in the quickly changing retail environment, tailored product recommendations are now crucial. While large retailers have employed market basket analysis (MBA) extensively, small, unstructured retail models such as kirana stores have not. This study explores the application of machine learning algorithms in generating customized product recommendations for a local supermarket. A transactional dataset comprising one month of sales records from a local supermarket (August 2024) was utilized for this study. The dataset contains 58,382 observations across nine variables. The objective of the study is to extract meaningful insights from the data by employing both association rule mining and clustering techniques. Specifically, the Apriori algorithm was applied to identify frequent itemset patterns, while the k-means clustering method was used to segment transactions into distinct groups for further analysis. The study produced 147 association rules, which could be used to create personalized product suggestions by offering insightful information about items purchased together. In addition, the k-means clustering algorithm segmented the transaction dataset into three optimal clusters, as determined by the silhouette coefficient metric. The managerial implications of these findings are discussed in detail. Specifically, the results indicate that personalized recommendation systems can substantially enhance customer satisfaction while simultaneously improving sales conversion rates. Furthermore, the implementation of such systems allows supermarkets to reduce excess inventory by aligning stock levels more closely with customer purchasing patterns, while also ensuring that shoppers receive relevant and timely product suggestions.