What is a ChatBot?
ChatBot, as the name suggests, is a robot made for chatting. ChatBot works by mimicking human beings such that the users feel that they are having a conversation with a real person. ChatBots have a wide range of applications in the customer care segments.
As the companies scale up, their number of customers also increases exponentially. It becomes physically impossible to deal with each customer manually at one point. Sometimes the problems are also repetitive & less complex. In such scenarios, using chatbots, we can suggest some elementary solutions to the user & forward more complex issues to the customer care agent.
A basic ChatBot that uses if-else statements is known as Command Bot. Command Bots generally give errors or default messages when a new/unseen command is entered.
A better & more complex version of ChatBot is a Learning Bot. Learning Bots uses Machine Learning and NLP algorithms to learn from the commands entered.
Amazon Lex
Source: aws.amazon.com
Amazon Lex is a tool provided by Amazon that helps us make chatbots on the AWS architecture. Companies can build customized chatbots with amazon lex and easily integrate them into their applications.
Amazon Alexa is also built using amazon's lex, and it uses the same underlying AWS architecture.
Amazon Lex chatbots use Natural Language Processing to convert a person’s speech into machine language & Natural Language Understanding to understand the intent or meaning of the text. The best part about amazon lex is that the developer does not require any machine learning, NLP, or Deep learning knowledge to build chatbots.
Use-Case of a ChatBot
Suppose a user wants to know his account balance without physically going to a branch. A conventional way would be to ask a customer care executive; the executive will manually authorize the person & then provide him with the account details, including the balance. With this method, the problem would be if, at the same time, many users call customer service to inquire about their balances; some will have to wait till an executive gets free to talk to them. The bank will also require to employ more customer care executives as they scale up.
A better solution for this problem would be to deploy a ChatBot on the Banking website. The customers can directly interact with the ChatBots to inquire about their banking details. The ChatBot will authenticate the customer and provide a set of choices to choose from. If the customer query is more complex, it can be re-directed to an execute.