Unanswered

Ai Builder Results shifting Data in Table

(0) Share

Report

Posted on by Community Power Pla...

I am trying to have AI Builder recognize certain information in payroll pdfs underlined in yellow, and place them into a table. I have erased most of the data from the payroll for security reasons. My table will have Name, Amount, MED, and Voucher Number (Check). The thing to note here is that CORY has no gross wages or MED, or voucher (check) number. So when I run a quick test, I would expect Corys name on the table would have blanks in the Gross Wages, Med, and Voucher Columns. BUT, what happens instead is it pulls the Gross Wages of Bobby up. Also combines check numbers from the employee below Bobby. See Quick test results below. What Can I do to remedy this? For the record 99% of the payroll DOESN'T have blanks like this, it only happens now and again.

Categories:

AI Builder

ExampleDoc2.png

ExampleDoc.png

I have the same question (0)

All responses (2)

Answers (0)

Varun8492 22 on at

Like (1)

Report
It seems like the AI Builder is not correctly recognizing the table structure in the payroll PDF, causing it to extract data incorrectly. Here are some suggestions that may help you remedy the issue:
Check the PDF file format: Ensure that the PDF files you are using for training the AI Builder are of good quality and are in a consistent format. Ensure that the underlines in yellow are consistent throughout all the files. If the underlines are not consistent, the AI Builder may not be able to accurately recognize them.
Modify the model to handle missing data: You can modify the AI Builder model to handle cases where some data may be missing. One way to do this is to add an "if" statement to your formula to check if the data field is empty, and if it is, replace it with a blank space. This will ensure that the data in the table is aligned correctly.
Train the AI Builder on more data: You may need to train the AI Builder on more examples of payroll PDFs with missing data. This will help the model better recognize the structure of the table and handle cases where data is missing.
Improve the table recognition: You can also try to improve the table recognition by modifying the AI Builder model. For example, you can try adjusting the table detection threshold or tweaking the layout options to improve the model's accuracy.
Overall, it may take some experimentation and tweaking to get the AI Builder model to accurately extract data from the payroll PDFs. With some adjustments, you should be able to achieve accurate results for most of the payroll data.

Was this reply helpful? Yes No
Community Power Pla... on at

Like (0)

Report

Thanks. This will give me new ideas to work with.

Was this reply helpful? Yes No