Before a chatbot can be realized which is able to talk in normal English, some kind of knowledgebase has to be created in the background which stores all the content. The problem is, that computers need a machine readable database which holds human-readable information and making this possible is one of the major tasks of Artificial Intelligence. A practical application of domain specific knowledge storage is a recommendation system, which can be used for a shopping website to guide the user to interesting products. If Fuzzy logic was utilized for converting the existing information into machine readable linguistic variables the recall rate isn't that great. The problem is, that only the average is built from different input features and the resulting output won't fit to the real needs of the user.
A similar matching problem becomes obvious, if medical information are stored in a semantic knowledgebase for creating an individual food recommendation system. The Fuzzy patient ontology contains the known information about blood sugar, age and heart rate of a person, and the rules are reasoning about the optimal meal the user has to eat. The Fuzzy logic rulebase will only make subjective decisions, the output of the system won't fit to objective criteria. The more reliable idea would be to use the Protego OWL editor without the Fuzzy addon, because this will allow the knowledge engineer to focus on the topic itself and ignore confusing non-boolean logic.