This can be optional. However, I chose to incorporate “speakable” JSON-LD for voice look for optimization, and I bundled the next snippet while in the schema markup: You will need to create your have JSON-LD script and spell out to the LLM what factors you wish to extract from the web