This is more about making a knowledgebase, or personality, for the actual chatbot engine. Weizenbaum’swas not only one of the earliest examples of a chatbot, but also of a “therapist” chatbot. I suggest to start by studying my webpage at . For instance, you could just use a copy of the original Eliza knowledgebase.
There are many examples of chatbots made from FAQs, or lists of frequently asked questions. Most chatbot knowledgebases are based on lists of question-answer pairs. There are also many examples of question answering systems based on question-answer pairs for various domains, or topic areas, for example library science, or librarian chatbots.
Most chatbot engines are based on markup language, ordialect, of one kind or another, often specific or proprietary to a given chatbot engine, or interpreter, the most common or paradigmatic chatbot language being . Commonly, chatbots are based on pattern matching, though there are examples of statistically based question answering systems. How terms or phrases are chosen and encoded for pattern matching is still as much art as science….
See also my Quora answers to: