Maybe I am confused but what does the AI do in your example? Isn't that just an instance of Google Maps itself being useful? What does the agent even spend its time doing?
The purpose of the demo is to show that you it's better to enable agents to access relevant data and actions in a controlled way than it is to do app specific integrations with, for example, a map and navigation app in order to plan group gatherings. Instead it's better to enable agents to access relevant data in a controlled way, and to leave coordination across apps to the agent.
In fewer words: if working with another app could be valuable, leave it up to the agent instead of doing the integration at the app level.