Speakable markup’s intended use is to allow applications and search engines the ability to identify content on web pages best suited to be read aloud by Google Assistant enabled devices. The benefit is the ability to reach a wider base of channels and users. 

With speakable markup, Google will produce the audio playback of the webpage, attributes the source, and sends the full article URL to a user’s mobile device through the Google Assistant App. 

This markup is currently only supported for English speakers in the United States and publishers that publish in English. 

What does your webpage need? 

1. cssSelector – Addresses content in the annotated pages (such as class attribute)

["headline", "summary"]

OR; But not BOTH 

2. xPath – Addresses content using xPaths (assuming an XML view of the content)


 Google’s Guidelines

For full documentation on speakable structured data visit: https://developers.google.com/search/docs/datatypes/speakable