Gmail's intent analysis magic behind separating emails into different mailboxes relies on natural language processing (NLP) techniques, which can also be applied to various other applications such as product success exploration, real-time transcription, and contextual insights. NLU is used for more dynamic voice interactions, creating a structure from unstructured data to define content. Implementing NLU using Python with tools like JohnSnowLabs' nlu and scikit-learn can help analyze text data from sources like Reddit reviews. Intent analysis has practical use cases such as client classification, intent marketing, spam detection, and modern uses in various industries, but also has limitations including incorrect results due to unstructured data and human behavior nuances. Tools like Symbl.ai offer custom trackers and APIs for building pipelines and gathering insights from voice and video data.