Showing results for 
Search instead for 
Did you mean: 
Frequent Visitor

How does form processing recognize text and tables


We are doing POC for customer on what Form Processing can do.

At the moment their sample documents contain a lot of text and tables (some spanning multiple pages as well).

Texts that we want to extract also don't really sit on the same location between documents. 
I know we can create document sets for different layouts, but in our case, they probably are too different with each other in terms of the content and tables.


How does AI Form Processing recognise tags and tables? 
Is it based on the location of the text/tables? Do documents need to have relatively similar locations of text/ tables and order?

For tables, do they need to have same headers? Can it recognise table based on tag outside the table e.g. Table title or caption?


Thank you.

I just couldn't find the exact documentation on how AI builder form processing works behind the scene.

Power Apps
Power Apps

If your documents don't differ too much with similar items location and table headers, they could be trained in the same collection.

There is some flexibility in the way the model will locate tables and their data to some degree.

Overwise, you probably have to build several collections trying to regroup documents which are somehow consistent.

To your question about tags outside the table it is difficult to answer without seeing samples of documents.

Best would be probably to try it out.

You can activate a one month trial to test the capabilities.

Helpful resources

Power Platform Conf 2022 768x460.jpg

Join us for Microsoft Power Platform Conference

The first Microsoft-sponsored Power Platform Conference is coming in September. 100+ speakers, 150+ sessions, and what's new and next for Power Platform.

365 EduCon 768x460.png

Microsoft 365 EduCon

Join us for two optional days of workshops and a 3-day conference, you can choose from over 130 sessions in multiple tracks and 25 workshops.

Top Solution Authors
Top Kudoed Authors
Users online (3,900)