What Is A Chatbot And How Does It Work

What is a Chatbot and how does it work

In recent times we cannot witness a website that offers services without a chatbot. Sometimes it is even annoying to see a chat window popping up for sales and marketing while gazing through the site. Chatbot is an offering typically used to start a conversation with users.

Chatbots use natural language processing (NLP) and set of pre-defined rules to understand the inputs as question and provide response closer to what is available with the knowledge database programmed in a much more interactive and intuitive way of communication.

For example, if I go to a bank website, I don’t have to eyeball through a galaxy of information available on the home page for a simple account creation. Instead, I can open the chatbot and ask the question “how to create an account”. It is one among the frequently asked questions and product information available somewhere within the website, but this assistance becomes a customer delight if someone going to get the information for me, isn’t it?!

What are the types of chatbots available presently?

Various chatbots are available, but most of them are broadly classified into the following categories.

Conversational Chatbots:

Bots intended to start a natural conversation with the users.

Task-oriented Chatbots:

Bots designed to help / assist users to complete a particular task or follow up a workflow until it reaches the completion state.

AI Powered Chatbots:

Bots which understands the user input and interpret the answer and then respond to the user. They are next generation of bots.

Most popular way of building chatbots

There are a dozen ways to create bots but some of them are very common that we already knew or even we use it for the daily driver, some of them are

Bot Framework: Microsoft introduces the Bot Framework which is a package of tools and services purposed for building the bots and consume via the line of the Microsoft products. For e.g., develop a bot using Bot Framework using Visual Studio Code, deploy in Azure by provisioning various services and users can consume via Microsoft Teams channel. Of course, these bots can be rendered for the outside of MS Teams as well.
	Slack APIs: Slack is very known, popular collaboration platform in the startups. Slack provides API which can be leveraged to build bots and can be integrated in the site to interact with the users in Slack Channels. Basically, the bots are meant to provide information messages, announcements and perform actions through Slack.
Telegram API: One among us reading this article would have used at least one of the bots available in Telegram. Even a couple of years before I thought of quitting Telegram because this is platform is encroached by bots which could be dangerous for my phone.. LOL… I understood wrongly. They are here to help us and not to spam or to attack the device. Of course, any creation intended for good action can be spoilt for one’s own benefit. Okay, recently I came across one bot which monitors an item in the ecommerce website and notifies when the price drops dramatically.
	Discord API: Now we are entering into a different universe, the world of gamers and other galaxies. Here bots are built for interacting with users in the Discord channels mainly. Live action during the gaming sessions, virtual meet-ups and announcements for community-based communications happens in Discord.

How does it work

The core functionality of a chatbot is driven by a set of algorithms and technologies that interprets user input, access and retrieve information from a knowledge database and provide responses. This algorithm works based out of combinations like

1. Natural Language Processing algorithms: To interpret and understand the intent behind the user message and generate responses.
2. Machine Learning algorithms: Track the user input and improve the performance of the automation over the time.
3. Knowledge based algorithms: Access and retrieve the information from a pre-defined set of information stored.

Two main techniques used to interpret the user inputs are

Syntax

Is an arrangement of words to break the sentence and segment it to parse and process the words.

Semantic Analysis

Is to understand the intent behind the message by recognition either it can be named entity or natural language generation.

What are the usages of chatbots

Chatbots can be used by a variety of businesses and services in a target specific format as specified below. But choosing which chatbot to be used will be decided by the capability. For example, rules based chatbots can be used for pre-defined set of rules to be executed and it cannot learn from the user interactions and provide responses for the open-ended questions.

Customer service chatbots

Virtual Assistant chatbots

Educational / Entertainment chatbots

News / Informational chatbots

\ \ Currently I'm planning to create this as a series to discuss in detail in the upcoming articles. Please stay tuned!