Product

The AI modules of Pridonu Cube

The modules can be combined with one another to suit your business process. Pick out the building blocks for your process – we are happy to help with questions or special requirements.

Module 1

Document separation

Any PDF stacks – for example all the attachments of an email, or several business transactions in one file – are analysed page by page. If an email contains several transactions, they become several separate cases, each with its own attachments. In each case the system recognises the hierarchy, across multiple levels too: a payment reminder covering several invoices, among them one invoice with an attached subcontractor invoice and its own delivery note. All without separator sheets, without barcodes, without pre-sorting. Type-specific extraction then follows for each recognised document; the separation results can be reviewed in the platform and corrected manually if needed.

Module 2

Extraction

Data from the document is read out by an in-house, multi-stage process. It uses the latest OCR techniques and AI methods. The structured data is transferred into a file and processed further. This extraction does not need to be trained – the system is ready to use straight away.

Module 3

Classification

The documents are classified automatically through layout and content analysis. You define which document categories exist – the system assigns them, or makes its own suggestion for classifying the incoming documents.

Module 4

Data enrichment / context

By linking various master data and data sources, the extracted data is verified and, where necessary, corrected automatically. The resulting data record is already checked and enriched with further information.

Module 5

Automatic checking

Using stored rule sets, information from the document is checked automatically – from complex calculations through duplicate checks to simple validation rules.

Module 6

Automatic rejection

Once all the information has been gathered, the document is rejected fully automatically – depending on how you define it. Because information about the sender is also available, no member of staff needs to step in.

Module 7

Post-processing

In a web-based system the document is displayed with all its information. If needed, the results can be reviewed or edited.

Module 8

Interpretation

The system interprets the document’s business context and “understands” the content of the data. This technique rests on a multi-stage AI rule system that can improve its own results quickly.

Module 9

AI suggestion

From all the gathered information, the system creates a suggestion for handling the matter. With this machine-generated suggestion, staff process documents considerably faster and to a higher standard.

Module 10

AI learning

The system learns in every area: in how the suggestions are used, in the handling of the document, and in the checking and interpretation of the matter.

Module 11

Reports

The system generates initial reports even during processing, so that you can actively intervene in the process. Once the business process is complete, additional analyses are available on request.

Which modules does your process need?

In a no-obligation conversation we clarify whether and how your document processes can be automated.

Book a demo Request a callback →+49 170 9650439