Automatic Document Recognition (IDR)

Disclaimer

All screenshots in the article were taken in the Dutch version of Yuki.

In Yuki, the processing of submitted documents is automated as much as possible. This applies not only to scanned documents, but also to files such as PDFs, Office files, and UBL/XML invoices.

Intelligent Document Recognition (IDR)

The process responsible for automatic processing is called Intelligent Document Recognition (IDR). IDR can classify, index, process, and validate documents. This automated intelligence is at the heart of YUKI RA (Robotic Accounting) and uses advanced AI-driven recognition software.

ATTENTION!

Soon, the recognition of Word documents (.doc) will no longer be supported. It is recommended to save invoices created in Word or Excel as PDF documents.

Triggers for Maximum recognition

If the IDR does not recognize one or more of the following fields on an invoice, the document is forwarded to Maximum recognition:

• Contact
• Document date / Invoice date
• Reference / Invoice number
• Amount
• VAT amount.

Maximum recognition is only available when a user with the ‘Portal administrator’ role has activated it for the domain in which the documents are processed.

For a detailed description of Maximum recognition, see article Activate Maximum recognition.

View IDR details

The specific details of the automatic processing of a document can be viewed by clicking on IDR details in the top right corner of a processed document.

Only users with the ‘Portal administrator’ or ‘Portal back office’ role can view the IDR details.

The following screen is opened:

The IDR details contain five tabs:

• Details
• XML
• OCR
• Enrichment (only available if Maximum recognition is activated for the domain)
• Log.
  
  

'Details' tab

This tab shows data on the:

• status of the job
• classification and default values of the document
• indexation of the fields in the document
• automatic processing of the document.

Status

The status of a job that consists of the recognition and processing of a document by the IDR and/or Maximum Recognition.

Classification and default values

Classification is the accurate recognition of the document type that has been submitted. The document type can be, for example, a purchase invoice, bank statement, pay slip, etc.

In addition, the result of the centrally (portal) and locally (domain) created processing rules that are applicable on the invoice is shown of the document type 'Purchase invoice'. 

Indexation

The indexing of a document consists of the dissection of a document and the recognition of specific fields in a document. This might include for example the recognition of a total amount from an invoice or the contact details of a supplier of an invoice.

The Recognition column shows the values in the fields as recognised by the IDR.

The Enrichment column displays the values in the fields as recognised or enriched by Maximum recognition.

The Actual Value column reflects the values in the fields at the time the invoice is handled automatically by Yuki or manually by the back office.

ATTENTION!

There is a red square in front of all fields not recognised by the IDR (Recognition column).

Automatic processing (completion)

The processing and automatic completion of the documents is based on the results obtained from the classification and indexing phase of the IDR process and Maximum recognition (if activated). 

Below Notes the reasons are stated why a document cannot be processed automatically. A few examples are:

• Invoices with an invoice amount above the 100.000 euro (and the document is no UBL).
  These invoices will always be offered to the back office for checking.
• The last 18 months no document is processed for the same administration/supplier/document type.
• The document is a duplicate.
• The document is part of a bundle.
• The supplier/customer is blocked manually for automatic processing (zie Define default values (processing rules) for supplier or customer).
• Etc.

Validation

After processing a document a decision can be made based on previous data whether or not the document is correctly processed automatically. In addition, a (quick) manual validation is carried out by an user with the 'Back office' role.

'XML' tab

This tab displays the results of the IDR. These results can also be displayed in XML or downloaded by clicking the XML or Download button, respectively.

'OCR' tab

Here you can see the results of the OCR.

OCR match rules

To improve recognition, users with the 'Back Office' role can create OCR match rules.

For a detailed description, see article Create local OCR match rule.

'Enrichment' tab

This tab displays the results of enrichment by Maximum Recognition. These results can also be displayed in XML by clicking the XML button.

'Log' tab

This is where a log is kept of all the steps taken to classify and index a document.

The recognition process is divided into different parts, also called funnels, corresponding to the fields in the input screen of a document.

Within each funnel, a number of checks are performed, which you can find in the log.