You’re offline. This is a read only version of the page.
I recently had a flow feedback loop engage using an AI model for document processing. When I went to the AI model to use the flagged file for training, I saw this message: "New documents from the Feedback loop are available for review. Add these documents in your training set and retrain the model to improve it." However, when I went to the appropriate document collection to add the document, I saw the below error - the file type is not supported.
Why isn't the PDF that originated the feedback loop engage listed? Have I missed a step in my feedback loop to make the entire document pull into this folder rather than just the input data? If so, how do I fix that? If not, maybe there's another error in the flow that I can't see? I did have the flow fail on this run, with data below.
I am using that action-- see below!
You need to use the action "Save file to AI Builder feedback loop". It will store the document in a Dataverse table. Just make sure that when you enter the source file in this action, you also add the file extension (.pdf for example).
Hi @Antrod thank you for the follow up! Is there a set location where I should be storing these files? I guess I assumed the feedback loop feature would take care of storing the document. Will the below work or do I need to add more steps to the flow?
Hi @Anonymous ,
It may be possible that within your flow, you didn't store the document with the extension suffix, example ".pdf". Can you double check that and let us know ?
The error you are getting in the flow is because you are comparing a value that is null, it's not directly related to the feedback loop process.
#1
WarrenBelz
791
Most Valuable Professional
#2
MS.Ragavendar
410
Super User 2025 Season 2
#3
mmbr1606
275
Super User 2025 Season 2
Share
Report
All responses (
Answers (