Covid 19 have accelerated the shift to online education from the traditional classroom teaching methods. The COVID-19 pandemic scenario, in specific, has brought out the need for innovative, creative approaches in the field of teaching. The Internet of Things (IoT) is a key enabling technology for creating intelligent structures that enable effective virtual and in-person learning environments. Since it has a direct impact on student engagement, enrolment, involvement, and in-depth knowledge, the move to smart learning—which incorporates IoT and Artificial Intelligence (AI)—into the educational system is enticing. Among the many problems affecting traditional education are teaching and learning, administration, assessment, and classroom supervision. Modern advancements in information and communication technology (ICT) have not adequately incorporated and sufficient into the framework of education. The present investigation looks at previous studies in the field and highlights over the following issues like problems with the current educational system and its possible solutions, the transition to smart learning and scientific obstacles (like cultural and technological barriers) to the transition to smart learning. Creative solutions to the problems with the traditional system, like smart administration, smart teaching, smart evaluation, and smart school environment has been discussed. This study suggests an innovative framework called the Reinforcement Learning based General Purpose online Learning Assessment Path Recommender System (RL-GPOLAPR) to determine the best solution for online learning assessment practices. Using the principles of reinforcement learning, RL-GPLAPR dynamically develops customized learning pathways according to each student's needs and choices.