Recently, someone asked me a question about Actions Builder on YouTube. The specific question isn’t very relevant, but I want to add my thoughts on this topic, and on voice bots in general.
I don’t use the Actions Builder, and I am going to wait for a while before using it.
Here is why:
It is pretty clear that Actions Builder is inspired by Dialogflow CX.
However, Actions Builder isn’t as powerful as Dialogflow CX.
Will there be a built in integration between Dialogflow CX and Actions on Google?
Now, technically, you can already combine the Actions SDK and use Dialogflow CX to power the NLU for your Google action. That’s a fairly laborious process though.
However, I am waiting to see if the Actions on Google team provides a direct integration between Google Actions and Dialogflow CX.
Even though Dialogflow CX started with almost no integrations, they recently added an integration with Dialogflow Messenger. It is entirely possible they might do the same for CX + Actions on Google.
Does it mean you should also wait?
No, it doesn’t.
You should proceed with Actions Builder if you want to get started building your Action right away.
Unfortunately, my view on the entire Actions on Google ecosystem right now is that the workflow from building the Action and eventually going all the way to deployment and app distribution is quite disorganized.
Some of it is because the tooling itself isn’t all that great. For example, Dialogflow ES is not bad, but it is very clear that Dialogflow CX will be a much better match for building AoG apps.
The other reason is that there does seem to be a little bit of neglect on the side of the AoG team too.
A lot of Google Actions went down (granted, more than a year back)
This is all my personal view, of course. But when someone posts a Reddit thread with the title “Does Google even care about developers?”, you (the reader) really need to start reconsidering whether you need to jump into the Actions on Google ecosystem.
After all, I could spend weeks and months on the process and the limitations of the service rather than my own technical skills could still prevent me from creating a successful app. And I am never going to get my time back. 🙂
If you are willing to take the risk (in terms of where you spend your time versus what the expected benefit could be), then you should certainly go for it.
But I am waiting to see if the overall process improves a little before jumping in to building AoG apps again.