Skip to main content

Google Assistant

Google Assistant is a voice assistant from Google that can be used on various devices.

Actions

Apart from the basic functionality, Google Assistant can execute commands from third-party developers. These commands are called Actions.

How Actions work

A user can open a bot using one of the activation phrases that can be set in the Actions console during the project registration.

When the user pronounces the activation phrase, the assistant switches to the dialog with the bot.

Limitations

Google Assistant has several limitations that you need to keep in mind when you create the script.

1. Use no more than two text answers at a time

You can add several bot’s remarks into one state. But the bot shouldn’t give more than two text responses at a time.

2. The picture must follow the text, not precede it

If your bot needs to show the pictures in its reply, put the picture after the text message.

3. Only one picture in the reply

You can only use one picture per reply. And, once again, it should go after the text message.

4. No more than 8 buttons in the reply

Google Assistant allows to use no more than eight buttons per reply. If you add more than 8 buttons to the reply, only the first eight buttons will be shown.

The design requirements

Google has certain design requirements for the conversational apps. If your chatbot does not follow those guidelines, your app will not pass the moderation and won’t be published in the Actions catalogue.

1. The bot should tell the users what action is expected from them

After the bot said its line, Google Assistant turns on the microphone and is expecting the reply. The user should understand what they need to say in response.

Add comments regarding the actions that you expect from the user.

2. Bot should be able to finish the dialog

Every bot in Google Assistant should know how to finish the dialog: to say goodbye to the user and turn off the microphone. Take this matter into account while developing the script.