One of the rapidly advancing applications in information technology is the recommendation system, designed to navigate the vast sea of available data and aid in decision-making. While numerous recommendation systems exist for various domains such as food, movies, and other commodities, they often rely on graphs, filters, and AI techniques, resulting in alternative results that lack clarity. This study endeavors to develop a recommendation system focused on selecting the most appropriate beneficiary. The universal struggle of individuals for life's essentials is a pervasive reality worldwide. Governments bear the responsibility of ensuring that welfare programs reach the populace seamlessly, promptly, and fairly, although this entails significant challenges in data generation, storage, and utilization. This research employs a Recurrent Neural Network, a deep learning model, to analyze and predict outcomes using crowdsourced data obtained from diverse government sectors directly connected with the public. The experiment is conducted using Keras and Tensorflow Python libraries, and the results are verified and compared with other recommendation systems. The RNN model achieves an accuracy of 93.01%, precision of 98.2%, recall of 94.3%, and F1-Score of 95.8% in the experiment. The problem statement addressed in this paper represents a novel contribution, and the achieved experimental accuracy is deemed satisfactory.