Chatbots: A new era that can transform Human – Machine Interaction

Chatbots: A new era that can transform Human – Machine Interaction

With the advent of AI and deep learning, it is evident that the so called “science fiction” concepts are transforming into a reality. Ranging from computer vision to exoplanet detection, AI has now become a de-facto standard in every business. Among the other applications of AI, the history of Human – Machine interaction has been evolving at a faster pace. From just typing commands to instruct through voice, it has come a long way and the latest addition is the Chatbots that uses natural conversation with a computer to instruct what you want it to do.

 

Is it a Robot?

While the term “bot” is usually visualized with a physical robot, it is not always the case. A Chatbot is a software application that is usually deployed in web fronts to interact with the users. It can also be voice enabled so that the users can interact through their voice. The users can interact with the Chatbots in their natural language (Commonly in English) and need not be grammatically correct and some sophisticated chatbots are now able to understand the common internet slangs and short forms of words too. (Can u lemme know ma acct balance? For example, can be understood). This brings in a whole new customer experience and instead of directing the user through complex screen navigations, A simple chat will do the necessary job for you.

The Magic behind Chatbot

Chatbots are powered by AI – Sounds too generic right? To narrow it down, Chatbots uses Natural Language Processing – A sub field of AI that can parse the user queries and extract the key information from it so that the computer can understand and perform the required task. Other Areas like Sentiment Analysis, Understanding the user intension (Natural Language Understanding) adds spice to the Chatbot and increases the engagement of the target customers. The beauty of Chatbots is that it keeps evolving based on the way the user interacts with them- the style, the tone , sentence formation – All gets studied and gets trained automatically so that the next time the user interacts, the chatbot will be able to provide a more accurate response with a personalized style for that particular user.

 

So How does a Chatbot work?

The user types in or speaks to the Chatbot interface – Be it a mobile chatbot app, web chatbot or voice assistants like Alexa, Google Home. The user’s query is sent to core chatbot engine comprising of NLP unit and Machine Learning unit. The NLP unit parses the user’s text or voice, tags them into different parts of speech (known as POS tagging) and identifies the key entities such as names, places, verbs, nouns etc. After identifying the key information, the machine learning unit maps them to different Intents – what the user is trying to ask, and then executes the respective business logic based on the identified intent. The pre trained language models developed using various machine learning techniques are heavily employed in Chatbots. These language models help to more accurately understand and interpret the user’s query to a machine-readable instruction.

So, in a nutshell, Chatbot is a combination of NLP and ML

 

 

Chatbot development frameworks

There are plenty of Chatbot development frameworks available from open source to commercial ones. These frameworks help to develop the chatbots quickly and on board them to your favorite deployment channels such as Facebook messenger, web, mobile, voice assistants like Google assistant, google home, Amazon Alexa etc.

So, you do not need to be an AI or Machine Learning expert to develop a Chatbot. Using the frameworks, you can design the conversation workflow using a visual development environment from less to no coding. Also, these frameworks support backend integration so that your chatbot can also talk to the database and bring in dynamic response to the user.

Some popular Chatbot development frameworks are Google Dialog flow, IBM Watson, Oracle Digital Assistants etc.

                 

Some of OneGlobe’s Chatbot Use cases

FAQ Bot:  Developed for a banking client and it addresses the common FAQ by the internal employees

Timesheet Bot:  Employees can chat and create their leave requests, weekly timesheets

Loan Bot: This bot can be used to verify the eligibility to lend Loans, get existing loan details etc.

 

  FAQ Bot:

Timesheet Bot: