Art of Communication for Chatbots

The key to effective communication does not lie in how great a speaker you are but in how good a listener you are. An amazing orator can have a monologue but it takes an attentive listener to have a delightful dialogue. It is only through a genuine attempt at understanding others, rather that being solely focused on conveying our beliefs, we can create the transcendental experience known to many as quality conversation.

There are several reasons active listening is an integral part of a good conversation. Firstly, we humans are often unsure of our own opinions and an attentive listener encourages us to delve deeper and find out an answer which we were, hitherto, unaware of. The process of exploring our thoughts and arriving at an answer is a satiating experience that only a good listener can provide. Moreover, stating your opinions very early on can, sometime, disengage a listener with a weakly formed differing opinion. A good listener often attains a greater understanding of the other person’s context and often delivers more apt and personalized responses, thus establishing a warm connection with his/her counterpart.

How can chatbots have better conversation?

  1. Focus on user understanding : Comprehending user request is the most important thing for the success of a chatbot. This is not a simple problem given that there are several ways of stating the same intent and two completely different intent might sound very similar. 
    I want my order” and “Where is my order?” mean the same thing but “I want my order” and “I want to order” are totally different.
    The ability of the bot-engine to differentiate between many such nuances is going to be a key UX differentiator.
  2. Provide feedback and ask questions: The bot should let user know when it is lost instead of throwing random responses. It should also be able to provide constructive feedback to drive the conversation .Based the previous conversations, the bot should prompt users to answer questions that can help it learn more about their preferences. Asking pertinent question and providing button-based navigation for them, can also help in eliminating NLP dependencies of the bot
    A female and a male user interested in buying formal attire should be asked different set of questions.
  3. Give personalized responses: The bot should be able to process user’s context tree and come up with responses. It is the quality and the relevance of these of these final responses that will that will determine end-user satisfactions. Most chatbots are going to be utility bots and if they are unable to carry out what they are designed to do, they are pretty much doomed.

Chatbots are soon going to be common place and will communicate with humans on a regular basis. For the success of these chatbots it is imperative that these interaction are productive as well as engaging. Mastering the art of communication is the new challenge that lies ahead of conversational interface designers who wish create exceptional chatbots.