Parixit     Jan,25 2018

Initiating IBM Watson AI Chatbot Development

These days building a ChatBot become more advanced without needing to code from scratch or doing hardcore programming by yourself. Chatbots have matured into solutions that enterprises across industries are taking seriously hence and no longer a nascent technology. Now, majority of the BOT Platforms provide advanced capabilities with predefined sets of variable objects, intents, entities, dialogs to setup as require and further can also do custom coding to meet expectations for exactly working ChatBOT.

 

IBM Watson is one the BOT development platform that provides capability to build AI Chatbot without doing hardcore programming. First to setup your company and region to avoid errors and click a little bit in the Web UI. Then follow these basic setup steps.

 

Below are the basic setup steps to build a ChatBot powered by IBM Watson conversation API.

 

IBM Watson conversation works only with below three objects :

1. Intents

2. Entities

3. Dialog.

 

Intents are the main user input, we can set goals or purpose. With the help of intents, one can start the conversation with Watson conversation BOT. 

 

Let's take an example to create sample conversation... In this example, we will see how to build conversation to Turn On/Off appliances of the house.

 

Based on our requirement we will create the intents #Intents. Here we have intention #TurnOn. 

#Intent

Here is the screen of intent #TurnOn

First is an intent name (Which will start with #).

Examples are the actual words or sentences by which Watson can identify the intention of the user. 

 

@Entity

The entity is the object which relates to your Intents. Here in our case Entity might represent the objects like Light, Music, Fan etc. 

As discussed, entity always prefixed by @ character.

It is good to set as many synonyms as you know to represent the entity. 

Like,

AC is also known as Cooling System, Cooling etc.

Music may have Radio, Music system etc.

Because a user can ask for the entity with any word which may relate or synonym of your entity.

As you can see in below image, I have set possible synonyms for my entity.

 

Dialog

A Dialog is the conversation flow that needs to define for your application to respond based on recognized Intent and Entity.

Dialog builder is a tool that may help you to manage the conversation flow. This will decide the response of the BOT.

You can set as many conditions as you want to set correct response.

Here is the image of dialog 

As you can see the tool of the dialog is the graphical representation of communication flow. It has condition block "If bot recognizes". 

You can set either Intent of Entity with "and"/"or" condition.

In our test case if TurnOn is not Music or 'AC' or 'Light' then send a response, "Sure! What you want to turn on?"

If either of the entity is available then we will redirect to next related dialog.

After this setup, it's time to test our chatbot.

Below image shows you the conversation between Watson BOT and User.

Here I am asking to turn on light. As we already set the intent to turn something "on" and set entity "Light" to that intent. 

Based on that Watson has responded "Ok! I am turning on Light".

If entity Light was not set as an entity it will ask to turn on something.

 

These are basic setup steps to initiate ChatBot Developement using IBM Watson Conversation AI.

 

Build, deploy and optimize chatbots quickly and efficiently with Watson Conversation’s advanced capabilities and easy to use interface. Get in touch with us!