12 Building for actions on Google with the Actions SDK
This chapter covers
- Differences between Dialogflow and Actions SDK actions
- The Actions SDK action package
- Analyzing text with regular expressions
- Using the Actions CLI
- Action deployment
Often in the book, we have examined the flow of information throughout the conversation. I’ve focused so much on this concept because I believe that by looking at this we can start to understand the system, understand what we can do, and understand where things might go wrong. In this chapter, we upend it—a bit.
12.1 Dialogflow and the Actions SDK
Let’s look again at the information flow in the Actions on Google app we built in the previous chapters, shown in figure 12.1. And look at the flow of a representative Alexa skill, shown in figure 12.2.
Figure 12.1 Data flows between user and fulfillment, with several steps in between.

Figure 12.2 Data flow in an Alexa skill

In both flows, among everything else, there is an NLU service. It’s very clear and discrete in the case of the Actions on Google app—the NLU is Dialogflow. The Alexa NLU is bundled up with everything else on the Alexa platform, but it’s still there.
That separateness on the part of the Actions on Google app opens up a question. Although there needs to be some level of NLU no matter what (no NLU, no conversational interface), can’t the developer own that? That’s what the Actions SDK provides, as you can see in figure 12.3.
Figure 12.3 Data flow in an Actions SDK app
