Building Solutions with the Microsoft Power Platform

Chapter 10. Continuing the Conversation on Chatbots

In Chapter 7, we got our first taste of using AI to create a basic conversational service to collect information with the help of a chatbot. In this chapter, we’re going to expand on the topic of chatbots and take a deeper look at Power Virtual Agents. The help desk bot that we created earlier was a simple example that showed how to collect information from a person who starts a conversation with the bot. In this chapter, we’ll discuss other features that make talking to our bots feel more like chatting with an actual person.

Using Prebuilt and Custom Entities

I introduced the concept of entities in Chapter 7; however, we did not make use of entities in the bot we built there. Entities are bits of information that you want to identify in a conversation. Common examples of entities include age, city, date, email, and money. If you configure your bot to request information for a specific type of entity, the bot will be able to use natural language to extract the appropriate information out of the sentence. Let’s take a look at a few examples of prebuilt Power Virtual Agents entities themselves, shown in Table 10-1.

Table 10-1. Examples of entities

Entity name	User input	Entity	Saved value
Age	I’m 19 years old	19 years	19
Age	My car is less than five years old	Five years	5
Age	She is a 10-month-old	10-month-old	10
Money	It costs 1000 Euros	1000 Euros	1000.00
Money	3 items for $1k	$1K	1000.00
Money	It costs a thousand ...