Integrate Dialogflow with Actions on Google

Summary

In this episode of 'Deconstructing Chatbots', host Priyanka Vergadia guides viewers through integrating Dialogflow with Actions on Google, enabling chatbots to reach over 500 million devices. The tutorial covers setting up explicit invocations, choosing intents, and testing the application. It also touches on user sign-in requirements, managing the assistant app, and preparing for deployment by selecting countries, surfaces, and providing necessary details for the app.

Takeaways

• 🤖 **Introduction to Integration**: The episode focuses on integrating Dialogflow with Google Assistant.
• 🔗 **One-Click Integration**: Dialogflow agents can be connected to multiple platforms with a one-click integration feature.
• 📱 **Reaching Devices**: Integration with Actions on Google allows access to over 500 million Google Assistant-enabled devices.
• 📅 **Example Bot**: The script uses an appointment scheduler chatbot as an example for integration.
• 🗣️ **Explicit Invocation**: Defines how the chatbot is invoked by a user's specific command, like 'OK, Google, talk to Smart Scheduler'.
• 🎯 **Intent Selection**: The Welcome Intent is chosen as the landing point for users after invocation.
• 🔐 **Sign-In Requirement**: Option to restrict service to signed-in users, but the demo keeps it open.
• 📝 **Testing**: Emphasizes the importance of testing the app before deployment.
• 🌐 **Global Reach**: During deployment, you can select countries and surfaces for your actions to run on.
• 📸 **App Details**: Requires a description, sample invocations, and images for the application.
• 🚀 **Deployment Options**: Offers alpha and beta release options before full production deployment.

Q & A

What is the purpose of the video 'Deconstructing Chatbots'?

The purpose of the video is to explore how to integrate Dialogflow with Google Assistant and other platforms.

What is Dialogflow and what can you do with it?

Dialogflow is a platform for building conversational agents. You can use it to create chatbots that can be integrated with various platforms like Google Assistant, Slack, Facebook Messenger, Twitter, etc.

What is Actions on Google and how does it relate to Dialogflow?

Actions on Google is a platform for creating software that extends the functionality of Google Assistant. By integrating Dialogflow with Actions on Google, you can reach a wide range of devices that support Google Assistant.

How many devices can you potentially reach by integrating Dialogflow with Actions on Google?

By integrating Dialogflow with Actions on Google, you can reach over 500 million devices that support Google Assistant.

What is an example of a device that supports Google Assistant?

Examples of devices that support Google Assistant include smart speakers, phones, cars, TVs, and watches.

What is the process for integrating Dialogflow with Google Assistant?

The process involves clicking on Integrations, selecting Google Assistant, defining Explicit Invocation, setting up the trigger name, and testing the integration.

What is Explicit Invocation and how is it set up in Dialogflow?

Explicit Invocation is when a user triggers an intent by saying a specific phrase, like 'OK, Google, talk to Smart Scheduler'. It is set up in Dialogflow by defining the intent to trigger when the user says the invocation phrase.

What is Implicit Invocation and how does it work?

Implicit Invocation allows the service to invoke an intent directly without the user needing to mention the invocation name, such as when a user says 'OK, Google, ask Smart Scheduler to set an appointment for 5:00 PM tomorrow'.

What is the Sign-In Required option and when would you use it?

The Sign-In Required option restricts the service to only signed-in users. You would use it if your service requires user authentication or personalized data.

How do you test your Dialogflow agent after integration with Google Assistant?

You test your Dialogflow agent by clicking on 'Test First' in the Dialogflow console to ensure it functions correctly after integration.

What is the Manage Assistant App button used for?

The Manage Assistant App button is used to set up the invocation name for your chatbot, select the voice gender for your assistant, and manage other settings related to the Google Assistant integration.

What are the steps to deploy your Dialogflow agent on Actions on Google?

To deploy your Dialogflow agent, you need to select the countries for deployment, choose the surfaces for your actions, provide a description, sample invocations, images, contact details, and privacy information, and then use the Release option to submit to production.

What is the difference between an alpha release and a beta release in Actions on Google?

An alpha release allows you to enable your action for up to 20 users without submitting for review, while a beta release opens it up to up to 200 users.

Outlines

00:00

🤖 Integrating Dialogflow with Google Assistant

Priyanka Vergadia introduces the episode on 'Deconstructing Chatbots' with a focus on integrating Dialogflow with Google Assistant. She explains that once an agent is built on Dialogflow, it can be connected to various platforms like Google Assistant, Slack, Facebook Messenger, and Twitter through one-click integration. The episode specifically looks at integrating with Actions on Google, which allows developers to create software that extends Google Assistant's functionality. By integrating Dialogflow, developers can reach over 500 million devices that support Google Assistant. The example of an appointment scheduler chatbot is used to illustrate the integration process. The steps include defining Explicit Invocation, setting up the trigger name, and choosing the Welcome Intent as the landing point for users. The process also covers implicit invocations, where specific intents can be invoked directly based on user requests. The episode touches on making the service available to signed-in users and testing the app before deployment.

Keywords

💡Dialogflow

Dialogflow is a Google-owned natural language understanding platform that allows developers to design conversational interfaces for websites, mobile applications, and various other digital platforms. In the context of the video, Dialogflow is used to build an agent which is then integrated with Google Assistant, enabling the chatbot to interact with users across multiple devices.

💡Google Assistant

Google Assistant is a virtual assistant developed by Google that offers a range of services, including answering questions, setting reminders, and controlling smart home devices. The video discusses how to integrate Dialogflow with Google Assistant, which allows chatbots to be accessed on over 500 million devices that support Google Assistant.

💡Actions on Google

Actions on Google is a platform that enables developers to create software extensions for Google Assistant. These 'actions' can be voice-activated apps that run on Google Assistant-enabled devices. The video explains how integrating Dialogflow with Actions on Google allows developers to reach a wide range of devices, enhancing the accessibility of their chatbots.

💡Integration

In the video, integration refers to the process of connecting the chatbot built on Dialogflow to Google Assistant and other platforms. This allows the chatbot to function across different devices and services, providing a seamless user experience. The script outlines steps for integrating Dialogflow with Google Assistant, emphasizing the ease of the one-click integration feature.

💡Smart Scheduler

Smart Scheduler is an example chatbot mentioned in the video designed to help users set appointments. It is used to illustrate the process of integration with Google Assistant. The video script describes how users can interact with Smart Scheduler through voice commands, demonstrating the practical application of integrating a chatbot with Google Assistant.

💡Explicit Invocation

Explicit Invocation is a term used in the context of voice-activated actions, where a user directly addresses the action by its invocation name. In the video, setting up an explicit invocation for the Smart Scheduler involves defining the intent that triggers when a user says, 'OK, Google, talk to Smart Scheduler.'

💡Implicit Invocation

Implicit Invocation occurs when a user's command is understood and acted upon without directly mentioning the action's name. The video gives an example of implicit invocation with the command, 'OK, Google, ask Smart Scheduler to set an appointment for 5:00 PM tomorrow,' which directly invokes the Schedule Appointment intent.

💡Sign-In Required

The 'Sign-In Required' option in the video refers to a setting that restricts the use of a chatbot to only those users who are signed in to their Google accounts. This can be useful for providing personalized services or ensuring user identity. The script mentions that for demonstration purposes, the Smart Scheduler chatbot is kept open to all users.

💡Test First

In the script, 'Test First' is a step where developers can test their chatbot's functionality before making it publicly available. This is crucial for ensuring that the chatbot works correctly and provides the intended user experience. The video emphasizes the importance of testing to avoid issues once the chatbot is deployed.

💡Manage Assistant App

The 'Manage Assistant App' button, as mentioned in the video, allows developers to configure various settings for their chatbot, such as setting the invocation name and choosing the voice for the assistant. This is part of the process of customizing the user interaction with the chatbot.

💡Deployment

Deployment in the video refers to the final stage where the chatbot is made available to users. Before deployment, developers must select the countries, surfaces (like smart speakers or phones), and provide necessary details like descriptions and images. The video highlights the importance of thorough preparation before submitting the chatbot for production.

Highlights

Introduction to 'Deconstructing Chatbots' with host Priyanka Vergadia.

Exploration of integrating Dialogflow with Google Assistant.

One-click integration allows connection to platforms like Google Assistant, Slack, and Facebook Messenger.

Actions on Google is a platform for creating software to extend Google Assistant's functionality.

Integration with Actions on Google reaches over 500 million devices.

Example of integrating an appointment scheduler chatbot with Actions on Google.

Instructions on navigating to Integrations and selecting Google Assistant.

Defining Explicit Invocation for the chatbot.

Setting up the Welcome Intent as the landing point for users.

Handling implicit invocations like setting an appointment.

Option to make the service available only for signed-in users.

Testing the app using the 'Test First' option.

Managing the Assistant App to set up the invocation name and voice.

Overview of preparing the application for deployment.

Selecting countries and surfaces for action deployment.

Providing a description, sample invocations, images, contact details, and privacy information for the application.

Instructions on using the Release option to submit the application to production.

Mention of enabling launch in beta for up to 200 users.

Option to enable alpha release for up to 20 users without review.

Summary of integrating the Appointment Scheduler chatbot with Actions on Google.

Teaser for the next episode to make the scheduler smarter.

Transcripts

PRIYANKA VERGADIA: Welcome to "Deconstructing Chatbots."

I am Priyanka Vergadia.

And in this episode, we are going

to explore how to integrate Dialogflow

with Google Assistant.

[MUSIC PLAYING]

Once you build your agent on Dialogflow,

you can use the one-click integration

to connect to various platforms, such as Google

Assistant, Slack, Facebook Messenger, Twitter, and more.

Today, we are going to look at integrating Dialogflow

with Actions on Google.

Actions on Google is a platform where

you can create software to extend the functionality

of Google Assistant.

By integrating Dialogflow with Actions on Google,

you can reach over 500 million devices

that support Google Assistant, such as smart speakers, phones,

cars, TVs, or watches.

Let's consider the same appointment scheduler chatbot

that we built in the previous episode and work

through the integration.

Click on Integrations.

We are going to explore Google Assistant integration today.

So let's just click that.

In the pop-up window, we define Explicit Invocation.

What intent do we want to trigger

when a user says, OK, Google, talk to Smart Scheduler?

Smart Scheduler here is our trigger name.

We will see where to set it up in just a little bit.

In our case, Welcome Intent is the perfect spot

for the user to land on.

So we will just stick to that.

Let's say a user says, OK, Google,

ask Smart Scheduler to set an appointment for 5:00

PM tomorrow.

That is an implicit invocation.

So in this case, we could invoke our Schedule Appointment intent

directly, since we know exactly what the user wants to do.

If we want the service to be only

available for signed-in users, then you

can check the Sign-In Required button.

In our case, it's just a demo, so we will just

keep it open to everybody.

We are then going to click on Test First.

Since we just developed our app, let's see how it functions,

or if it functions correctly.

OK, that looks great.

Now let's explore the Manage Assistant App button.

This provides us to set up the invocation name to our chatbot.

We called it Smart Scheduler.

This is also the place to select male or female voice

for your assistant.

Going back to Overview, we have already built our application.

And it's ready to test.

When you are ready for deployment,

make sure you have selected the countries you want to deploy in

and the surfaces you want your actions to run on.

You will have to provide a description

for your application, Sample Invocations.

For example, here we've added Talk to Smart Scheduler.

You can add more.

It also requires two images--

one small and one large.

Contact details and privacy information

fill up all those things.

Once you have done all that and feel

comfortable to push the application to production,

use the Release option to submit to production.

Of course, here, we have not finished the overviewed items.

So we don't have the ability to submit to production yet.

You can also enable the launch in beta

to open it up for up to 200 users.

With alpha release, you can enable your action

for up to 20 users without submitting for review.

All right.

So today, we used our Appointment Scheduler chatbot

that we built in the previous episode

and integrated it with Actions on Google.

Don't miss the next episode of deconstructing chatbots,

because we are going to make our appointment

scheduler a little bit smarter.

[MUSIC PLAYING]