Exercise 2 - Test the small talk dialog from IBM Watson Assistant
The Small Talk service has been implemented using only the IBM Watson content (i.e. Intents, Entities and Dialogs). The content exists in both the IT and HR Chatbot. The Small Talk content was developed in the IT Chatbot and then the Intent and Entities where exported and manually imported into the HR Chatbot. The Watson Dialog Copy Tool, documented in the guide, was used to copy the Small Talk Dialog Nodes from the IT Chatbot to the HR Chatbot.
Intents:
There are a total of 23 intents that come with the Small Talk content. All intent names start with #BMC_SmallTalk_ and match the Small Talk content list above.
From these Intents, four of them use the concept of Annotation. These are:
- BMC_SmallTalk_DoYouHate, with their annotations stored in the entity @SmallTalk_DoYouHate
- BMC_SmallTalk_DoYouLike, with their annotations stored in the entities; @SmallTalk_DoYouEnjoy, @SmallTalk_DoYouLike and @SmallTalk_DoYouLove (Note: this intent has three entities created as Watson was having issue with the annotations when they where in one entity collection)
- BMC_SmallTalk_Other_AreYou, with their annotations stored in the entity @SmallTalk_Other_AreYou
- BMC_SmallTalk_Other_Asks, with their annotations stored in the entity @SmallTalk_Other_Asks
The annotated word is typically the last word or words in the sentence. For example, in the sentence, "Do you hate working", "working" is annotated so that the bot can use other words in its place to derive the same response. In the intent the annotated word is highlighted within the intent value as shown below: 
For more information regarding annotations please refer to the following link: https://cloud.ibm.com/docs/assistant?topic=assistant-entities#entities-create-annotation-based
Note: for a complete list of utterances used in the various small talk intents, please review them using the IBM Watson workspace/skill where these intents have been installed.
Entities:
There are a total of 47 entities that come with the Small Talk content. All entity names start with @SmallTalk_. Most of the entities created for small talk are used to control the selection of the Small Talk Dialog Nodes. They are typically used in conjunction with the Intent to ensure that the Small Talk dialog is selected with greater confidence as both the intent and actual conversation snippets, which are stored in the entities, have been used in the user's utterance.
Note: for a complete list of entities, please review them using the IBM Watson workspace/skill where these entities have been installed.
Dialogs:
There are a total of 26 parent dialog nodes that come with the Small Talk content. All the dialog nodes have been placed on a parent folder called BMC - Small Talk so that they can be more readily found within the overall Dialog nodes structure for the bot. This is shown below: 
We will not document all the Dialog logs nodes here. Please review them using the IBM Watson workspace/skill where these dialogs have been installed. We will review a few examples here in order to illustrate some of the interesting implementation concepts.
Dialog Node Examples:
- As mentioned earlier some of the parent dialog nodes use both an Intent and an Entity with their "If assistant recognizes" condition so that the Bot will only run the dialog node if both are true. For example, the BMC - Small Talk – ChatBotDataIntent patent node will only run if it recognizes the intent utterances in the #BMC_SmallTalk_ChatBotDataIntent intent and the conversation snippets are used from the @SmallTalk_ChatBotDataIntent entity.
- Within the BMC - Small Talk - Feedback – Abusive dialog node we have used the concept of a counter to determine how many times the user has activities this node. As the count increases, the user gets different message content as well as different options (in the form of buttons) to choose from. For example: the dialog has been setup to display three different messages and options based on the users activating the parent node the first time, then the second time and finally the third and subsequent times.
The first node initiate the counter to 1. Every time the user activates this node, the counter is increased by one as the content variable Feedback_A is not re-initialized in the session. 
For example, when the counter reaches 3 or higher, the user gets a different message and a number of options such as quit the session or transfer to a live agent.
- Addition of Easter Eggs that can be found only if you play around with the bot. For example, one of the answers provided by the bot to the question "Where do you live" is "I live in the Matrix." If the user asks about the Matrix, then they will be given a fun response from the movie. For example, if the user asks "Are we in the Matrix" then the dialog node BMC - Small Talk - Fun_Matrix will run.
There are several Easter Eggs that come with the Small Talk content. The purpose is again to provide a more social and fun aspect of the Bot to try and get users to return. Updating this type of content over time also helps. - Teach the Chatbot to tell jokes and have fun. The dialog node BMC - SmallTalk_General_Joke has several jokes that have been added for the Chatbot to use. Again, this content can be updated over time to keep it fresh.Also, the Chatbot has been trained to get the user a cup of coffee or tea as seen from the dialog node BMC - Small Talk - Other - Asks - coffee/tea