Availability of content and training sets is a major bottleneck for a chatbot development today. Relying on Apache OpenNLP and its sub-project OpenNLP.chatbot, we introduce a number of tools and components to design a chatbot and its training set to be knowledgeable and intelligent. n In this talk we will analyze the reasons it is so hard to find a chatbot demo today for a nontrivial task or to observe an intelligent behavior of a chatbot. It is easy to see how a success in AI can boost the chatbot development on one hand, but it is hard to detect intelligence in those chatbots that are available to the public, on the other hand. n We will present an advanced search engine for chatbots with the focus on linguistic features and discourse-level analysis for dialogue management. We will introduce a tool that builds a dialogue from an arbitrary document to form a training dataset for deep learning chatbots. We will demo a chatbot supporting virtual dialogue, where a user joins a virtual community built on the fly, whose members answer questions in this user’s current area of interest. An extended content for this talk is available in the book recently published by the speaker “Developing Enterprise Chatbots”.