This document discusses adding natural language understanding capabilities to mobile apps. It covers concepts like intent detection, entity extraction, and context. It also provides an overview of tools that can be used to build conversational interfaces like Dialogflow for intent training and testing, Wit.ai, Core ML, and TF-IDF algorithms. Challenges with multi-language support, offline vs online capabilities, imperfect understanding, user adoption, and whether to add voice to apps are also discussed.