Extract Rows and Columns from Word Tables

(1) Share

Report

Posted on by Eric_B

Is it possible to extract row and column data from a table in a Word document using Power Automate? I’m aware that Word supports inserting tables (Insert → Table) and typing data into the cells of the table, but I don’t see any Power Automate actions that allow me to read or extract data from an existing table in a Word file. Is there a built-in action, connector capability, or workaround that I may be unaware of?

Categories:

Building flows

I have the same question (0)

All responses (5)

Answers (0)

Sort by

Suggested answer

11manish 2,252 on at

Like (1)

Report
Copy link

Link copied!
Power Automate does not natively support extracting table data from Word documents.

To achieve it, you must use:

Content Controls (best structured approach)

or AI Builder / conversion-based workarounds

Was this reply helpful? Yes No
Suggested answer

Valantis 5,108 on at

Like (1)

Report
Copy link

Link copied!

Hi @EB-04061852-0,

No built-in Power Automate cloud flow action reads tables from Word directly. The Word Online connector only populates templates and converts files.

Here are the confirmed options:

1. AI Builder document processing (cloud flow): train a model on your Word documents and it extracts table rows and columns as structured data. Each table row is returned in an Apply to each loop. Requires AI Builder credits but is the easiest cloud-native path.
Ref: https://learn.microsoft.com/en-us/ai-builder/form-processing-model-in-flow

2. Power Automate Desktop + VBScript (desktop flow): Microsoft docs confirm you can use the Run VBScript action to open a Word file via COM and read its content including table cells. Requires PAD and a machine to run on.
Ref: https://learn.microsoft.com/en-us/power-automate/desktop-flows/how-to/extract-data-from-word-table

3. Parse the docx as XML (cloud flow): a Word .docx file is a zip containing XML. You can get the file content, extract the document.xml, and use xpath expressions to pull table cell values. No extra licensing but requires xpath knowledge.
Ref: https://tachytelic.net/2021/05/power-automate-extract-text-from-word-docx-file/

4. Azure AI Document Intelligence: best for complex or inconsistent table layouts, returns precise row/column data. Requires Azure subscription.

What does your table look like and are you building a cloud flow or desktop flow? That determines which option fits.

Best regards,

Valantis

✅ If this helped solve your issue, please Accept as Solution so others can find it quickly.

❤️ If it didn’t fully solve it but was still useful, please click “Yes” on “Was this reply helpful?” or leave a Like :).

🏷️ For follow-ups @Valantis.

📝 https://valantisond365.com/

💼 LinkedIn

▶️ YouTube

Was this reply helpful? Yes No
Eric_B 47 on at

Like (1)

Report
Copy link

Link copied!

Valantis -I am building a cloud flow using the SharePoint trigger “When a file is created or modified (properties only).” The flow targets a Word document stored in a SharePoint Document Library and is triggered based on a column named Effective Date.

The Word document contains a table that was likely created using Insert → Table. The table includes four columns: Revision, Date, Description of Changes, and Document Owner. However, the header names and column structure may not be consistent across documents. The table is typically located near the bottom of the document.

Let me know if you need further information. - Eric

Was this reply helpful? Yes No
Suggested answer

Valantis 5,108 on at

Like (1)

Report
Copy link

Link copied!

Hi @EB-04061852-0,

Thanks for the details Eric. Given your setup cloud flow, SharePoint trigger, inconsistent headers this narrows it down to two realistic options.

AI Builder is out because the inconsistent column headers mean you can't reliably train a model on it. PAD is out because you're on a cloud flow.

Option 1: Parse the docx as XML (no extra licensing)

Since it's a cloud flow you can get the file content directly from SharePoint, then extract the table rows by parsing the underlying Open XML. The steps:

1. Get file content using path (SharePoint)
2. The .docx is a zip — use a Compose action to work with the raw XML or call an Azure Function to unzip and extract the document.xml
3. In document.xml, tables are in `<w:tbl>` elements, rows in `<w:tr>`, cells in `<w:tc>`, and text in `<w:t>`
4. Use the xpath expression in a Parse XML action to pull out cell values row by row

This works well for a table at the bottom of the document even with inconsistent headers since you're reading raw cell positions not named columns.

Option 2: Azure AI Document Intelligence (most reliable for your case)

Since the headers aren't consistent, Document Intelligence's prebuilt Layout model is actually the better fit than a custom-trained AI Builder model. It reads tables by position rather than header name, returns each cell with its row and column index, and works on any Word or PDF file without training. Requires an

Azure subscription but the pay-as-you-go pricing is low for occasional document processing.

For a one-time or low-volume trigger like a SharePoint modified file, the XML approach is the cheapest. For anything more than a few documents per day,

Document Intelligence is cleaner and more maintainable.

Which direction fits your volume and Azure access?

Best regards,

Valantis

✅ If this helped solve your issue, please Accept as Solution so others can find it quickly.

❤️ If it didn’t fully solve it but was still useful, please click “Yes” on “Was this reply helpful?” or leave a Like :).

🏷️ For follow-ups @Valantis.

📝 https://valantisond365.com/

💼 LinkedIn

▶️ YouTube

Was this reply helpful? Yes No
Eric_B 47 on at

Like (1)

Report
Copy link

Link copied!

@Valantis - Your right, it will depend on the volume of documents that get released; we could have as much as five documents go out on a given day, then we wouldn't have any for days, even weeks, at a time.

The Azure AI Document Intelligence looks promising based on what you shared, but I am not against using the XML Parser either. I guess it is just what flavor I want. What do you recommend? Is there documentation on either process?

Was this reply helpful? Yes No