Usually chatbots doesn’t know nothing at the beginning. You have to learn them to answer on visitors’ questions. SecondEGO provides some nice features to help you learning your chatbot. One of them are predefined knowledge modules.
Internally, all modules are extremely carefully structured and bigger than you expect. Only this approach allows developers to catch visitors’ questions. They deliberately follow for possible awkwardness which can shows over conversations and edit the modules if necessary.
Descriptions for knowledge modules:
Personality – replies on questions about chatbots’ personality (name, identity, sex, age, location, mood, profession, interests…).
General knowledge – replies on usual and frequently questions (hour, date, calculation, jokes, geography, weather…).
Wikipedia – replies on questions when chatbot expect answer in Wikipedia (what is, who is…). Here is the example of one contemporary question about ebola:
Fillers – fillers are general replies to unknown questions. We can explain unknown questions as visitors’ question where chatbot can’t find an answer within his set of answers (site search, patterns, previous knowledge modules…). In this case chatbot will use some special replies named fillers. This can be an anaphoric questions (why, when, how, examples, where…) or some sporadic questions or even replies from language corpus of varied articles or documents.
As I said, predefined knowledge modules in SecondEGO are very nice addition to start with your chatbot. I would suggest that in some cases (for example in business application) you don’t enable Chat and (especially) Fillers modules. Fillers can sometimes alter conversation to different level and it is great that you can simple disable module. But for starters and personal usage all of available modules are fine.