We have implemented a Chatbot using IBM Watson conversation service to answer home insurance related questions. Here is what we did:

1) Log into IBM Bluemix using your login credentials. The Overview screen appears.
Note: If you are a first time user, create an account in IBM Bluemix and after login, you will be asked to create organization and space. After creating them you will see a bunch of Bluemix cloud based services.

2) Click on Watson category.

3) Click on Get started now and select a conversation from the list of Watson services.

4) Provide a name in the Service Name field and click on Create button as shown at the bottom of the page.

5) Click on the Launch tool in the service dashboard. Upon prompt, create a workspace by clicking Create. Provide a name, description and select the language which you want to use in the conversation. After creating the workspace click on it.

I want my Chatbot to answer the following questions:

  1. What is home insurance?
  2. How much coverage do I need?
  3. What deductible should I choose?
  4. What is my home insurance in Boston?

Adding Intents and Examples:

An intent is the purpose or goal of a user's input and Examples are utterances that fit a particular intent. So, to recognize an intent we need to add as many examples as possible.

1) Add intents by clicking on Add Intent.

The above questions will be answered by the following Intents.

Here are the Intents and their associated examples:

Intent : #defineterms:


  • I want to know what home insurance is
  • Define home insurance deductible
  • What is my coverage

Intent: #location_insurance:


  • What could be my home insurance in Boston
  • How much does home insurance cost in Boston

2) Adding Entities:

An entity is a portion of the user's input that you can use to provide a different response or action to a particular intent. You can add entities in entities tab.

The entities I have created for the intents are as follows:

Intent: @related_terms:


  • Coverage
  • Deductible
  • Home Insurance

Intent: @location


  • Boston
  • Chicago
  • Los Angeles

3) Providing Dialog:

The Dialog component identifies the Intent and Entities, and provides a response to the user. The Dialog contains boxes known as nodes. Each node will have a condition and text that should be provided to the user. You can create child nodes to ask more information based on the user’s input.

Anything else node is used whenever conditions are not met by the user’s input.

The conversation bot will be automatically trained. So, you can test your bot at Try it out.