Rethinking customer support automation is necessary as businesses often over-automate tasks that humans are better suited for, leading to frustration and resentment from customers. The key to success lies in understanding when customers have common + easy questions or rare + hard questions, and using computers to answer the former while involving humans for the latter. A great support product should understand this distinction, make it fast to resolve common questions, provide a way for customers to give feedback on answers, and allow them to move from helping themselves to talking with human support staff when needed. An AI-powered solution like Articles aims to achieve this by using machine learning to suggest content to both humans and customers, enabling self-service and efficient resolution of common questions while involving humans for rare and hard queries.