Rise of Chatbots

This is an era of AI (Artificial Intelligence) and as an industry we are rapidly adopting the AI technologies into the businesses to go more faster, make more profits while increasing the customer satisfaction and engagement. One such popular adoption is a ChatBot !!

According to the surveys conducted, 90% of the apps installed are used only once and people spend more time on messaging apps than on social networking sites, making the messaging platforms new social networks today !! Businesses and brands have sensed it and moved to the messaging platforms to host the ChatBots to be right there, where the customers are.

Messenger, WeChat, Skype, Slack and bunch of other messaging apps are now hosting ChatBots on their platforms, with all this you as a customer  can now chat with CNN to get the news or to book the flowers for your Bae or order your favourite pizza by chatting on your messaging application that you use a dozen times a day and you do not need to download any other app, the bot resides inside the messaging application you are using.

This is just a beginning of ChatBots and are not saturated yet being in an era of early adoption. Introducing the Deep learning has got all the new age and brains to the Bots now unlike before they were just designed as an answering machines with hard coded decision making based on the inputs. Just in the couple of years we have seen the Chatbots becoming an integral part of the messaging apps from being just an Automated FAQ applications in the past. There is an evolution happening at a very high pace.


Everyone of us today use a chatbot in one of the other form, be it Google Assistant on android platforms or Siri on iOS to the home automation solutions like Google Home and Alexa. We are using bots as assistants, booking flights, ordering food, customer service, HR, sell insurance policies, beauty tips, health tips and medicines and so on to just to name a few. In short, bots are everywhere and we rely on them to a certain extent, this makes it extremely important to assure the quality of the chatbots and test them thoroughly. 

Let us understand what why and how of chatbots before we look into the Test Strategy for Chatbots.

What is a ChatBot?

Think of a Chatbot as a messaging tool between a human and a robot, which uses Artificial Intelligence and natural language processing to interact with the humans. 

Chatbots are broadly classified into 2 types:

Rule Based:
These bots are designed based on some presets of inputs with the responses. Based on the user inputs the decision making algorithm would match a input and output set.

AI Based:
These bots come with their own brains and ability to continuously learn on their own based on the conversations they have. They also analyze the sentiments and can have a small talk and keep the customer engaged. They are implemented using Deep Leaning algorithms and Natural Language processing.


How are ChatBots helping the Businesses?
Today the users expect the services to be available 24x7 on their finger tips with no delays with  personalized support. Not all the businesses can afford the staff to be available across 24 hours a day. Chatbots are one such solutions which can be implemented by the businesses to gain benefits. Some of the benefits they provide would be,

  • Personalization
  •  24*7 availability, Bots never sleep
  •  More Customer Engagement
  • Cost Effective
  • Less training and faster on-boarding
  • Alternative Sales Channel

With all the pros there are certain concerns that ChatBots especially the AI based Chatbots bring on to the table. Let us see one of such example.

Tay - Microsoft Bot

Tay was a chatter bot released on Twitter by Microsoft which they had to shut down within 16 hours of its release when it learnt abusing  based on its conversations. You can read more here.

Poncho, the popular weather app

Bots that tries to do too much, has been sending users messages unrelated to weather.

One cannot ignore adverse effect of AI implementations and still need to put in a lot of efforts on thoughts on how can we control them. Through testing is definitely one of the keys here. 

In the next article we will see how are the chatbots different from any other applications and what are the challenges in testing them

Comments

Post a Comment

Popular posts from this blog

Cross Browser testing for React Apps

Before we discuss the need of cross browser testing for React apps lets get a quick essence of why do people use react to develop their apps and what is cross browser testing. React React and React Native , support all the major browsers across platforms ability. It means that if one would want to build an application for iOS and android platform, all they would do is have a single code base and rest is taken care by react to make it cross platform. In addition to it the support for all the major browsers and has been tested, however with a little limitations for older browser where the polyfills are used to support them. Cross browser Testing Cross browser testing helps you validate if your application is compatible across different browsers, platforms and devices. It becomes important today to support all the major browsers and the devices as it is almost impractical to control what platform or browsers the users would use to access your application i
Read more

Fixing the appium installation with NPM

Problem Statement: npm install -g appium returns the below     npm install -g appium  npm ERR! code ENOENT  npm ERR! syscall chmod  npm ERR! path /usr/local/lib/node_modules/appium/build/lib/main.js  npm ERR! errno -2 Solution The directory npm is trying to modify is not owned by the user that is executing the command. Grant the permissions to the user executing the command and the issue should be resolved.  Here are the commands:  sudo chown -R `whoami` ~/.npm sudo chown -R `whoami` /usr/local/lib/node_module Uninstall the appium npm uninstall -g appium Re-install appium npm install -g appium
Read more

Challenges in testing ChatBots

        This is the second article on ChatBot Testing. Here is the first article which introduces the chatbots in general and details out why should we test chatbots.         Chatbots are different to other applications in several ways as listed below which makes them challenging to test and measure the quality. Self Learning Systems         Chatbots are built using AI and NLP services and keep learning and improving on their own. This makes it highly unpredictable whether a test case has failed due to the defect or if its improvement. One cannot have the traditional style of testing and write test cases with expected output. Remember, the test data used itself can be used by the bot to learn and update its behaviour, the negative testing may have an impact on the bot !! Non-linear input          Voice based chatbots use the speech recognition technologies to convert the voice to text. Chatbots are heavily dependent one such solutions. Along with the dependency comes
Read more