This guest column is by Amit Arora, CEO, Miracle Group
A chatbot is a program meant to mimic human responses and interact with people as a human would.
Tay, a chatbot attached to an artificial intelligence developed by Microsoft’s Technology and Research team and Bing search engine team was not only designed to do just that, it learned about human conversation from the internet by interacting with people on Twitter. However, within 24 hours, Tay had endorsed Donald Trump, claimed Hitler was right and made a slew of misogynistic comments. Since then, Microsoft has disconnected Tay and deleted a number of her inappropriate tweets.
Chatbots, however, are nothing new. A realistic conversational computer program has been a goal of
artificial intelligence research since Alan Turing proposed his famous test in 1950. They are as old as chatrooms, including the bulletin board systems that dominated the internet before the invention and adoption of the web. The SmarterChild bot on AOL Instant Messenger and MSN Messenger (now Windows Live Messenger) had 10 million active users and processed 1 billion messages per day. It provided access to news, weather, sports, a personal assistant, calculator etc.
Interested in building your own? There are a number of ways to build a chatbot but the easiest method would be to use existing NLP (natural language processing) engines. Below are some options to help get you started.
This is an advanced artificial intelligence chatbot platform and living community of chatbots and people. Personality Forge’s award-winning engine integrates memories, emotions, knowledge of thousands of words, sentence structure, unmatched pattern-matching capabilities, and scripting language called AIScript. The language is easy to work with and requires no prior programming experience to get started.
The AI engine uses both NLP and Case-Based Reasoning (CBR) which are two philosophies of AI. In NLP, sentences are parsed and broken down to reveal the structure of the sentence and information about individual words and their relation to other words in the sentence. In Case-Based Reasoning, sentences are searched for key phrases which trigger pre-programmed responses. Performing both, the AI engine first breaks down the sentences using NLP into their most basic elements, finds relationships between those elements, finds the meaning of individual words, and then passes all this information forward to the key phrase, or CBR section. Responses are matched against both specific and broad categories of statements, and then the response is constructed using both the bot’s own original words and a wealth of information available from the other chatter’s message and memories of the other chatter.
The Personality Forge’s own scripting language, AIScript, takes this flexibility even further by providing the ability to create if-statements and responses based on memories, emotion, sex, time, and date. Many famous chatbots that have won chatterbox challenges and other contests have been created using Personality Forge. Creating an account is free and can be done in minutes.
Pandorabots is another platform for building and deploying chatbots. The Pandorabots Playground is free with an integrated development environment for building chatbots loaded with features and tools.
The platform offers tutorials for those looking to hone their skills or brush up on the basics of building chatbots. You can view and download conversation logs to use during bot development and integrate your bot into applications using AlaaS.
You can create ALICE/AIML based bots, edit the response and publish. If you don’t want to learn AIML and build your own chatbot, Pandorabots’ team of engineers and AIML experts can build a virtual agent according to your needs and specifications and can even help integrate avatar or speech capabilities.
You can also sign up and access Pandorabots’ key features for free.
Facebook Messenger hopes to make it simple for developers to reach people on mobile, no matter how small the company or idea and gradually, Facebook will be reviewing submissions of bots built on the platform to ensure the best experiences for everyone on Messenger.
Within bots for Messenger, you can send and receive API which includes the ability to send and receive text, images, and rich bubbles with CTAs. Facebook has also created generic message templates with call to actions, horizontal scroll, urls, and postbacks. Other capabilities with Facebook Messenger are welcome screen and Null state CTAs where users can see your brand, your Messenger greeting, and a call to action to “Get Started” with the idea in mind to let users run the experience.
Messenger has also deployed a tool enabling conversational bots that can automatically chat with users. Making it easier for users to discover your bots, Messenger has made it possible to deploy Web plugins, Messenger Codes, Messenger Links, or Messenger Usernames.
For more information and how to get started building with Facebook Messenger, you can visit
There are also a number of step by step guides and online tutorials for those who want to build their own chatbot from scratch. This is complicated however especially if you want your chatbot to be truly capable of interpreting what the user is actually saying and then generating an appropriate response.
Hopefully, the options listed in this article are less overwhelming and will help you complete your first project in no time.
Image Credit – Shutterstock