Skip to main content

Tracked Changes in .docx Files

Using Hatz Document Editing

Updated this week

What Hatz means when it talks about using tracked changes:

"We can programmatically generate a Word document that includes 'suggested edits' (redlines) by manipulating the raw code of the file, allowing you to see exactly what we changed versus the original text."

"Advanced Edits" refers to using OOXML (Office Open XML) to make tracked changes or redlining in Word documents - basically the "Track Changes" feature you see in Microsoft Word where edits are highlighted and can be accepted/rejected.

How to prompt for this:

You can ask something like "Add review comments to the paragraphs that mention pricing" or request tracked changes when editing text. The AI will apply these edits using the underlying OOXML format that preserves the change tracking.

Use cases:

- Document review workflows where you need to show what was changed

- Collaborative editing where multiple people need to see proposed changes

- Legal or compliance documents where change history matters

- Any situation where you want reviewers to approve/reject specific edits rather than just getting a final version

This works with .docx files and maintains the professional review process you'd expect in business document editing.

AI can work with existing comments in .docx files. It can read and interact with comments that are already in your document.

What you can do:

- Ask the AI to respond to existing comments

- Have it add new review comments to specific sections

- Request it to address or incorporate feedback from existing comments into the document text

Limitations to keep in mind:

- The AI works with the OOXML structure, so it handles comments as data objects rather than having a visual interface like Word

- Complex comment threads or nested replies might not be processed as smoothly as simple comments

- When the AI processes your document, some formatting or comment positioning could shift slightly during the editing process

You could prompt something like "Review the existing comments in this document and incorporate the suggested changes into the main text" or "Respond to each comment with a brief acknowledgment and make the requested edits."

Disclaimer: AI is non-deterministic by nature, so using AI generated outputs to resolve comments will still require human oversight.

Did this answer your question?