A.I., Chatbot and Me .. Learn how to develop a chatbot that fits your business

Chatbot (A chat  ro bot ) : A program developed through computer, made to simulate an intelligent conversation or chat, through "Ar...

Chatbot (A chat robot ) : A program developed through computer, made to simulate an intelligent conversation or chat, through "Artificial Intelligence" over an instant messaging interface.

Some of the well know chatbots : Apple's Siri, Microsoft's Cortana, Google's Assistant, Google Allo, Amazon's Alexa, Facebook's Messenger. All these are different versions of chatbots. Beyond these branded chatbots there are good number of other chatbots integrated across various sectors ranging from customer service, checking weather to doing more complicated and logical tasks. In 2016 alone the no. of chatbots developed were somewhere around 34000 and increasing. As this is a leading technology and a trend changer, many AI based startups were founded and later some of them are being merged or acquired by giant tech companies and here is the report.

Pic Credits: CBINSIGHTS 

Technically what makes a bot to chat and be a chatbot ?

A bot can be a good one or a bad one. The developers intention behind the development of a bot to perform some task is the one that characterises the bot. I hope you will not enjoy reading details on bad bots,  so here I start with A.I..

Artificial Intelligence : A concept developed as program with abilities to solve problems on it's own for specified tasks or posses human like intelligence for visual perception, decision making and for speech recognition. Computer on it's own cannot solve or do such tasks as and when you ask it to do, it needs to have AI powered program. AI powered programs can also be integrated into website's and instant messengers (chatbots).

AI for website : Is a mix of UI and Artificial Intelligence to have a well designed and programmed user interface. Example's: Slack, Weps and The Grid .. the known AI Websites and Website Builders. 

Molly - The Grid's AI Web Designer

Chatbots : These are being used across industries and for various purposes. These are categorised into AI based and Non-AI based.

Non-A.I. based chatbots are scripted chatbots, these basically can't do any task which needed intelligence.

A.I. based chatbots are classified based on their value proposed. These are designed and developed to perform specified task, or provide services on demand. Example providing weather information or news. Based on nature of task they perform they are classified into different types.

"Chatbots are designed to accomplish a task".


Social bots: Social Chatbots run based on unique of a group or crowd available online in the social media's messaging platforms, such as Facebook's Messenger, Slack bots, Tinder Stacks and few more.

Chatty's : Chatty bots (chatbots) designed just to have conversation with users (website visitors). This is to make user's engage on website, which indirectly helps to accomplish social and commercial goals.

Shield's : Shield bot, a chatbot to execute customer support activities, comparing to feature their abilities and benefits than others and also includes fighting a parking ticket. For this, probably the best example and easy to recognise is the "Chatbot Lawyer". These kind of chatbots are also called as optimiser chatbots built to provide pleasant and friendly user experiences.

Optimiser's :  These optimiser chatbots are largely available functional bots designated to assist users in getting results for the query they raised and concentrated on providing better user experience.

Pro-active's : These bots act according to time and place to provide right information at right time and are specialised in it. They provide notifications without much disturbing the user.

One-Trick Pony's : One thing at a time, doing mini activities as an utility tool built into messaging platforms which can convert images to meme's or add text to videos. These are simply excellent bots which can reflect cognitive capabilities.

Finally here we come, to the well known and other amazing chatbots available in the market :

"The Raise of Digital Assistants"

Super bots : Google Allo & Assistant, Microsoft Cortana, Amazon Alexa, Apple Siri and the upcoming Samsung Bixby.  These are the intelligent personal assistants powered to perform various actions and show up their skills. Some of these are built into  operating systems available across multiple platforms and some into digital devices.  Hoping many of you are aware of these chatbots as apps on your mobile / desktop and digital devices.

Our digital assistants/chatbots are still not smarter why ? Do we need to speak in their understandable language or will they have to understand humans natural language. Here comes the importance of NLP ( Natural Language Processing ) which helps in enabling Conversational A.I.

When NLP comes to discussion, one might hear of NLU ( Natural Language Understanding ) being used as interchangeable term with  NLP.  NLP allows the systems to work together and handle end to end interactions between human and machine in the human's preferred language. To keep it simple if show you the difference between NLP and NLU.

NLP : When a query is raised by user, NLP accepts the query, breaks it down to comprehend it's meaning to take appropriate action and respond as quickly as possible in user understandable language.

NLU : Deals with complex challenges of A.I., that is to handle unstructured inputs governed by flexible rules for converting them into structured format, that a machine can understand. NLU is just about the machine's ability to understand what user says and act accordingly.  Statistical Modelling, Sentiment Analysis and conversion or filtering of speech or text i.e., Computational Linguistics come into picture when you talk about NLU.  Example : Seeker (An AI SalesMan).

If you like to try having fun with bots , then go with Instant Translator, Chillin, Chatty, Vexera and for Data Science enthusiasts try Innodatatics_bot over Telegram app. 

Other amazing bots & cognitive agents:  

Techcrunch : Helps users to get latest news through the subscriptions they took from various topics.

Trim : This helps in clearing out or unsubscribing from various subscriptions user took, by analysing transaction details.

Instalocate : A flight tracking chatbot integrated with Facebook messenger.

Letstock : A chatbot to get recent stock quotes.

Seeker  (An AI SalesMan) from SeekNShop.io :

An upcoming pluggable cognitive agent for e-commerce platforms with intuitive chat and voice interfaces. This allows end users to discover products along with contextual product recommendation and real time promotions. Which is based on engagement of customers on sellers web/mobile channels, provides real-time insights for sellers.

Shoppers prefer physical store experience due to human interaction, assistance and support offered by customer service representative's. The common customer shopping pattern's such as bargaining, requiring a guided product discovery and recommendations to avoid confusion, makes a customer to potential buyer and a next time visitor as well. But this isn't possible in e-commerce just because of the sheer scale to be catered to millions of users. Providing a similar support or shopping experience as in-store shopping isn't possible without AI.

Seeker made it possible. It re-creates best in-store experience by leveraging Artificial Intelligence and the power of live conversations with customers, by using some patent pending techniques involving deep learning and reinforcement learning in digital/ online shopping platforms. And this is a perfect blend of technology i.e., NLU + Cognitive Assistance which learns from customer interactions and cross references to several user conversations to respond better and to ease customers shopping experience. The opportunities for cognitive agents in business are unexploited. However it's time they have made their way into the market with AI.

Here are some reference's which we personally collected, to help you find details on :

Pic Credits Quora.com

Artificial Intelligence, Chatbots, Machine Learning, Deep Learning are trending now and spreading across various industries like Medical, Education, Automobile, IT, FinTech, Banking, HR, Manufacturing. This is an untapped opportunity for people to build something worthy by developing applications that truly services the customer.  

How can I choose or develop the best bot that fits my business ?

You need to follow these steps to develop a chatbot

1. What are the goals you want achieve by building a chatbot
2. Identify the opportunities you want to capture by using a A.I. based chatbot
3. Design chatbot conversations
4. Use frameworks or development platforms to build the chatbot

Along with these, businesses need to identify what work can be automated and what complexity can be reduced by using Artificial Intelligence solutions. By analysing the activities one can see  that complexity can be in data or work. These activities are intended to Innovate, Effective, Efficient and be as an Expert in providing solutions. 

Here are few known chatbot 

  • builder platforms : Motion AI, Converse.io, Octane.ai . 
  • developer frameworks : Microsoft bot framework, API.ai , Wit.ai (Facebook's bot engine)
Choose the required and make the best out of it. 

Chatbots can be used for automating routine tasks without loosing the charm of your business. Use human intelligence for productive activities.

 If you are eager enough to know about acquirers of A.I based startups then look at CBINSIGHTS infographic as per July 2017.


You Might Also Like



Flickr Images