Natural language understanding (NLU) is an important domain within human-computer interaction that enables machines to comprehend and interpret human language. This chapter explores how reinforcement learning (RL) can be applied to enhance NLU systems, addressing long-standing challenges and opening up new possibilities. From extracting meaning in text-based games to personalizing user interactions, we’ll examine cutting-edge techniques that are pushing the boundaries of what’s possible in NLU. We’ll also investigate how NLU can, in turn, inform and improve RL algorithms, creating a symbiotic relationship between these two powerful technologies.