@JO443 ,It is hard to advise without looking at your source and output documents.
But the general idea is that once you get the JSON data, like below
[
"How to extract ",
"text ",
"from a Microsoft Word docx file",
".",
"This document explains how to extract text from a Microsoft Word document using standard Power Automate actions. The result isn’t perfect, but it should be good enough for basic usage."
]
you could then use the Populate Word Template action to populate the relevant fields by parsing the array variable after some string manipulation.
How To Populate a Word Document Template with Power Automate (citizendevelopmentacademy.com)
The option suggested by @Mira_Ghaly is also a good one. It is useful if you have documents that are of similar format (such as invoices) that could be used to train a recognition model, which can extract relevant content by comparing your document against the model. You need a separate AI builder license for it though.