Extract data of pdf with Aibuilder

(0) Share

Report

Posted on by Try

Hi i want to extract pdf data with Ai builder in power automate desktop anybody know solution please help

Categories:

Power Automate Desktop

I have the same question (0)

All responses (4)

Answers (0)

Sort by

Sundeep_Malik 6,501 on at

Like
a
(0)

Report
Copy link

Link copied!

Hey @Try

I dont think AI Builder is available for Power Automate Desktop.
So, I will suggest you some workarounds.

Save the file from local to Sharepoint library, then just call a cloud flow with http request that will pick the file from sharepoint, and extract the data, then if you want to perform further steps in Power Automate Desktop you can send the data back to Desktop flow or trigger another flow.

Second method is to use one of the following:

Rest tagging AI Builder experts, if they have any other solution. There could be an API that could invoke AI Builder directly from Power Automate Desktop.

@JoeF-MSFT @CedrickB @plarrue @Antrod

Was this reply helpful? Yes No
momlo 1,527 Super User 2024 Season 1 on at

Like
a
(1)

Report
Copy link

Link copied!
As @Sundeep_Malik wrote, currently there are no AI Builder actions built into the PAD.
SharePoint related actions were the first ones and I heard from MS rep for Power Platform in my country they were expecting more to come sooner but got some delays/problems with that.

Do you really need a AI Builder to accomplish your tasks?
If not - try to see if built-in PDF actions work for you if your files have text layer, and/or OCR actions if they are scanned images.

Then if you still need AI Builder there are 2 ways:
Mentioned by @Sundeep_Malik - save file on SharePoint, and build online flow that runs for newly created files, extract data from file, store it in SharePoint, get data back to PAD.
Downside is you need to build logic.
What I have build in one of my solutions:
A. Bot 1 - Get files from system A - save them to SPO library via folder synced with OneDrive
B. Cloud flow - Run for all new pdf files in a folder - extract data, do some pretty complex logic with extracted data, delete file. At the end add the extracted data to excel table
C. Bot 2 - wait until all files are gone from folder (meaning all cloud flows are completed). Get list from excel file and process in system A

What is good with this approach is performance:
A. Bot 1 - is simple, has 1 task - save file
B. Uploading to SPO is done by OneDrive so it is faster then building this into Bot 1 and processing files in a sequence with built in actions
C. Cloud flows run simultaneously, so again - they one file is not blocking other ones
D. Bot 2 - quite simple too, has 1 task - read excel is and process data
Another approach would be to create cloud flow that is called with API, and you pass the PDF content in the api call, cloud flow does the job with AIB and pass back the results, process the result.
I did not go with this route as this would force me to process the files sequentially* so performance would suffer.

*there is still room to make it simultaneous, by storing the cloud flows results in a excel file.
But that would make it very similar to solution 1 so I just settled with 1 🙂

Was this reply helpful? Yes No
Try 33 on at

Like
a
(0)

Report
Copy link

Link copied!

Hi i created cloud flow i got response and how to transfer that response to pad

Was this reply helpful? Yes No
VJR 7,635 on at

Like
a
(1)

Report
Copy link

Link copied!

Take a look at the below video if it suits your purpose.

Advanced Invoice Processing with AI Builder in Power Automate Desktop - YouTube

Was this reply helpful? Yes No