IDP - Process cant recognize language/words in Reconcile form but extracts correct data (pic inside)

Hello,

I stumbled upon a little problem with IDP.

When I extract data from Appians built in IDP, the document is processed nicely, I can see the document clearly. But when I created a process from tutorial:

 https://docs.appian.com/suite/help/22.4/create-doc-extraction-process.html

 I can see something like this in Reconcile form:

It looks like Appian does not recognize language, except 2 words, and whats stranger for me, extracted data is fine, both from auto mapping and from manual extract.

What may be causing this situation and whats more important, what can be the solution? 

I'm trying to skip reconcile part, but I would like to show to my supervisor some results and I want to get rid off the view like in screenshot above.

Also one more question about IDP and its "learning". 

Does IDP learn only when it writes data to database? In the built in IDP, when i uploaded few files, there is a progress with how well he captures data. Each new document gave better results.

But when I tried it upload a document few times for my simplier process from tutorial, it doesnt seem to "get smarter". Is it due to lack of saving data or maybe built in Appians IDP has more Nodes in process that tell him to learn? I'm just trying to wrap my head around this and understand it well.

Thank You for your time.

  Discussion posts and replies are publicly visible