I recently added the Dialogflow Messenger integration for CX on my web app (which helps you generate Dialogflow FAQ bots with a single click).
Here is a little screenshot:
I have already removed it from my web app though, and I will explain why later.
Here are some tips for people who are just getting started with this integration.
How to enable Dialogflow Messenger integration for Dialogflow CX
Here are the steps
1 Each CX bot has an associated Google Cloud project
Inside your Google Cloud Console, you can select the project ID from the top menu.
2 You need to enable billing for the given project
When you sign up for the free trial, you would have associated a credit card with your Google cloud account. But you still need to enable billing for the specific CX project.
3 You need to enable Dialogflow API for the project
For the project associated with the CX bot, you also need to enable the Dialogflow API.
4 Inside your Dialogflow CX console, click on Manage tab
Click on the Manage tab to see the Integrations available
5 Choose the Dialogflow Messenger integration
Click on Integrations from the left hand menu, and click on Dialogflow Messenger integration.
6 In the popup dialog, click on Try It Now button
When you click on the Try It Now button, you should be able to interact with your bot if you set everything up correctly.
If your bot works inside the console, it will also work on the web page where you embed it.
I also noticed a few more things with the integration.
No chat prompt
If you look at the Dialogflow Messenger integrations based on ES on my website, you will see that they provide a chat prompt automatically without the user typing anything. This doesn’t work with your CX bot as of now.
No history of user conversations
This was the reason I decided to remove the bot from my website. There is no History or log of conversations, and as of this writing, there is no way to know what the user said to the bot.
Interestingly, there are some conversation analytics which show summarized stats, organized by Intents. But how do you know if the summary is correct, if you cannot be sure that the user’s phrase actually got mapped to the correct intent?
This is the same issue I raised in my article about why I don’t recommend Chatbase to my clients.
I have mentioned this before, but Dialogflow CX is still a little undercooked, in my personal opinion. While it is a good idea to start learning Dialogflow CX now, you might want to wait for a while before creating production bots based on CX.
- Actions Builder vs Dialogflow CX
- 5+ ways Dialogflow CX is better than Dialogflow ES
- How to bulk upload training phrases for Dialogflow Messenger
- Dialogflow CX vs ES: First look
- How to send rich responses from webhook to Dialogflow Messenger
- Dialogflow CX now generally available
- Understanding Dialogflow CX Parameters
- Dialogflow CX Missing Features
- Dialogflow Messenger integration for CX: First look
- Dialogflow Conversation Analytics Tips