In this tutorial, I explain how to add a chatbot to your website using the Dialogflow Messenger integration.
What we will be building
The video below shows what we will be building in this tutorial.
Embed the chat widget into your website
Embedding the chat widget into your website is straight-forward.
Create your Dialogflow agent
Create a new Dialogflow agent, and remove the Welcome event from the Default Welcome Intent.
Add a new intent into your agent with the following information (note that it includes the Welcome event).
Add the script to your web page
Once you create your welcome intent, go to the Integrations pane, and select the Dialogflow Messenger integration.
Now copy the script into your clipboard.
Paste the script you copied into the <head> tag of your HTML page.
In that case, use a plugin like Insert Headers and Footers and add the script into the “Scripts in Header” section.
WordPress Gutenberg Update
With the release of Gutenberg, you can now simply add HTML code blocks directly into a page. So add a HTML block in Gutenberg and paste the script you copied from the Dialogflow console directly into the editor.
Test the integration
When you open the page in your browser, you will now see the chat widget at the bottom right of your web page.
Now click on the welcome message, and you should see the below.
HTML and CSS customizations
The big advantage of using the Dialogflow Messenger integration (instead of the 1-click web demo) is that it is possible to customize the behavior of the chat widget using HTML and CSS customizations. You can check out the different customizations available in the documentation, and I also go over them in more detail in the Dialogflow Messenger Bots course.
For example, change the chat-title in your script within the df-messenger tag to “Dialogflow Messenger Bot”.
When you open the page, you will see the new title in your chat window. If you don’t provide any value for the chat-title attribute, the name of the Dialogflow agent will be used as default.
By far, the biggest feature of the Dialogflow Messenger integration is its support for rich responses such as links and images and buttons.
Let us go over these rich responses one by one.
The description type allows you to add multiple lines of text under a heading.
Create an intent called show.description.type as below: To see the result, type the appropriate phrase into the chat widget.
Info Response type
The info type is similar to the card in the Google Assistant rich response, and it supports an image, a title and some text below the title.
Create a new intent in your Dialogflow agent called show.info.type. The image URL points to an image on my website. Now check the result in the chat widget. You should be able to click anywhere inside the Info response type and a new browser tab will open and navigate to Google.com
You can also display images inside the web chat window in the Dialogflow Messenger. However, the image auto-resizes to fit the width of the window, so small images get expanded and look distorted. Larger images generally look better.
Photo credit: Unsplash
Add a new intent called show.image.type in your Dialogflow agent as below. Now test the response in the chat widget.
Buttons are a good way to add a link to some page on your website. I don’t recommend using the “event” parameter in the button type, and I explain why in the Dialogflow Messenger Bots course.
Create an intent called show.button.type as below.
Now test it in your chat widget.
Suggestion chips are good choices for representing buttons in your Dialogflow chat widget. I don’t recommend using it for links, though, and I explain why in the Dialogflow Messenger Bots course.
Create an intent called show.suggestion.chip.type as below. Now test it inside your chat widget.
Note: the custom payload which is generated by the tool is exactly what you need for the suggestion chips to behave like buttons, as described here.
That is, it doesn’t support all the fields used in the suggestion chips, precisely because we need the functionality (of being able to click on the suggestion and have it submit as text inside the chat window).
Tip: You can use the BotFlo app to automatically generate the custom payload JSON for suggestion chips.
List Response Type
The list response type is an interesting one.
Compared to the other rich responses, it is more complex to implement. But it also provides some benefits you don’t have in other rich responses. Specifically, if you need a “listbox” style input – where the presented value has to be different from the underlying key – it is your only choice.
You need to add a second intent into your agent with the event PARALLEL as below.
Remember that you already have an intent called df-messenger-welcome with another event – the Welcome event.
Now add another intent in your agent as below:
Now you can test it in your chat widget. When you click on the second list item, it will trigger the parallel universe intent.
The accordion type displays collapsible information.
Add another intent to your Dialogflow agent as below.
Now test it in your chat widget.
Do you notice something unusual?
Yes, the Accordion supports HTML content! 🙂
"Can I do this in Dialogflow Messenger?" I am often asked about these things, which you cannot do in Dialogflow Messenger, or at least not without significant programming cost - add bot typing delay - allow the visitor to maintain the conversation session across multiple pages - reprompt the user automatically after some period of inactivity - automatically end the conversation after prolonged user inactivity - transfer chat to live agent - automatically get email notification when chat completed - automatically save user information to a database If you need any of these things, I recommend a low cost alternative called Zoho SalesIQ. See how SalesIQ compares with Dialogflow Messenger. You can integrate SalesIQ with Dialogflow ES with the click of a button.
Sending Rich responses from webhook code
This website contains affiliate links. See the disclosure page for more details.
"The magic key I needed as a non-programmer" The custom payload generator was the magic key I needed (as a non-programmer) to build a good demo with rich responses in DialogFlow Messenger. I've only used it for 30 minutes and am thrilled. I've spent hours trying to figure out some of the intricacies of DialogFlow on my own. Over and over, I kept coming back to Aravind's tutorials available on-line. I trust the other functionalities I learn to use in the app will save me additional time and heartburn. - Kathleen R Cofounder, gathrHealth