Now that you have your Dialogflow agent and fulfillment deployed, you can test out your Action on real hardware devices or the Actions Console simulator.
In addition to thorough testing, make sure your Action passes our pre-launch checklists before submitting. This ensures a quicker approval and catches a lot of the issues we see during the approval process.
Testing on real hardware
On a supported device, such as a voice-activated speaker or an Android phone, log into the device with the same account that you used to create your Actions on Google project.
Test your Action in the Actions Console simulator
- Click Integrations in Dialogflow's navigation menu.
Click Google Assistant and click TEST in the pop-up window. Dialogflow uploads the agent to our servers so you can test it in the Actions simulator.
In the Actions Simulator, enter
Talk to my test appin the Input area of the simulator to test your Action.
Submit your Action for approval
Dialogflow versioning allows you to take a snapshot of your agent to create a version that is immutable. You can create as many versions as you want and deploy a specific version to production or alpha/beta environments.
With the Dialogflow versioning feature, any edits you make to your Dialogflow agent will not impact your publicly available Actions. You can iterate and improve your Action in draft mode and make it available to users only when you're ready.
You’ll also be able to release different versions of your Actions in alpha, beta, or production release environments and roll back to previous versions if necessary.
Publishing to alpha/beta release environments
Dialogflow versioning is enabled automatically for Actions, when you submit an Action that you built with Dialogflow through the Actions Console.
To submit Actions from the Dialogflow Console to a Actions on Google release environment, follow these steps:
- Click Integrations in the navigation menu, click Google Assistant and, click MANAGE ASSISTANT APP in the pop-up window.
- In the Actions Console, follow the steps described in Set up alpha/beta releases.
How does Dialogflow versioning impact my existing Actions?
- Existing Actions submitted to the Actions Console prior to May 8, 2018 will not be impacted. This means that edits you make to your Dialogflow agent will continue to be reflected in the existing Action. If the Action is already published, the edits will continue to be reflected live.
Will there be two versions I need to maintain in Action on Google and Dialogflow?
- No, you have only one version as shown in your Manage releases page. It is unified across Actions on Google and Dialogflow. The only difference is now that we will create a snapshot of your Dialogflow agent at the time of submission. This snapshot will not be impacted by any future edits you make to your Dialogflow intents and entities.
Can I create and deploy Actions on Google versions from the Dialogflow Console?
- No, you can only do so via the Actions Console.
In the Dialogflow page I saw that this is available for users who opt-in to beta, is that applicable for Actions on Google?
- No, the Dialogflow beta program is for Dialogflow agents that are not on the Actions on Google platform. If you have a Dialogflow agent for the Actions on Google platform, then it is automatically enabled and not in beta.
What happens if I enable beta in the Dialogflow Console?
- Doing so will not change Actions on Google projects. If you enable beta, you will see an Environment tab in the Dialogflow Agent settings. In that page, you can view (but not edit) the same versions you created from the Actions Console and load any specific ones into the draft.
What is the suggested best practice for using alpha/beta/production releases with Cloud Functions for Firebase?
- Dialogflow's inline editor does not have versioning for Cloud Functions for Firebase. However, each Dialogflow version points to custom webhook parameters. If you want to test different server logic with Alpha/Beta releases, we recommend that you turn off the inline editor, deploy your solutions into two Cloud Functions and pass those as webhook parameters in Dialogflow.