IBM AI Training Chatbot -Part 1
The first step towards setting up a chatbot is to identify the purpose of the chatbot. The hypothesis that I want to test is the efficacy of a chatbot as a training tool and hence this is the purpose of my chatbot.
An IBM Chatbot requires the below minimum setup to build a conversational Bot.
- IBM Cloud Account: Create a free IBM Cloud account and add the resource ‘Watson Assistant’
- Assistant: An assistant helps the end users with their requests directed to the Bot. I created an assistant with the name ‘AWS Certification Assistant’
- Skills: Skills contain the training (that we need to configure) so that the Bot (Assistant) can respond to user queries
- Every Assistant needs to be associated with one skill at the least.
- A Skill can be a Dialog Skill or Actions Skill. We will look at details of each type of skill during actual setup
- Intents: In very basic terms, Intents are the reasons why an end-user may interact with the bot. E.g. book a hotel or get account information etc.
- Intents are defined within the Skill setup.
For the bot to identify an intent, we need to setup user examples for each intent.
E.g. for an Intent ‘#Cancel_Account”, below can be sample examples:
- Cancel my account
- I want to cancel
- Close my account
- I want to drop my account
- I want to cancel my policy
Dialog: Simply put, setting up a Dialog is the process of configuring the responses to the user queries.
User questions (assigned to Intents) are routed to the relevant response- text, option based response, image, transfer to a channel or human agent.
With this basic understanding, we will configure the Bot in the next posts.
