“Templateless Invoice Extraction” Self Learning

Hi ,
I tried to create “Templateless Invoice Extraction” (sample project attached) but sometime the PO number is not extracted and I have to open “Present Validation Station” to add the correct value.

Please let me know what can I do in order to make the project extract the PO number when the invoice template is the same.

Thanks,
Rafik Beshara

1 Like

Hello @Rafik.Beshara

Welcome to the community!!!

I have also come across this same scenario multiple times through Templateless invoice extraction. What I did is, I used a Regex extractor along with the machine learning extractor to match the regex pattern and capture the fields that machine learning extractor is unable to capture directly. I think this same approach can help get your scenario fixed too…

Try it out and let me know how it goes…

5 Likes

Hi @Lahiru.Fernando

Thanks for your replay.

How can I use Regex extractor in this case, the pattern is shared ,for example PO# is number(digit) and invoice# is number (digit) so how can I write Regex extractor to get PO# only.

Thanks,
Rafik Beshara

where is “Templateless Invoice Extraction” (sample project ?

Project is presented on a screen shot, it the first post, here: Link

hi pawel,

With “Configure Extractors” link within the Data Extraction Scope activity - if i try to select and match any label with ML extractor vs taxonomy then it shows API key error. Please let me know what is the problem?

hi, have you found the solution yet?

NO, nobody is telling anything.